Open Hub
Open Hub

 Settings |  Report Duplicate

0
I Use This!
Very Low Activity

Commits : Listings

Analyzed about 7 hours ago. based on code collected 1 day ago.
Nov 05, 2025 — Nov 05, 2026
Commit Message Contributor Files Modified Lines Added Lines Removed Code Location Date
+--------------------------------------------------------------------------------------------------------------------------------------------- +1252. (THEORY and FEATURE) Merit of Large Scale Visuals - GIS Urban Sprawl Analytics - Automatic Delineation of Urban Growth Boundaries from Transportation Network GIS - related to 1151,1174,1196,1226,1233,1234,1237 and all sections on GIS Urban Sprawl Analytics, Merit of Large Scal e Visuals, Gridded Population Estimators (SEDAC GPW4 and GHSL convex hull estimators), DBSCAN and Neural Network Urban Extent Learners, Comput ational Geometry, Intrinsic Merit Rankings, R Ricker Population Estimation, Community detection, Social Network Analytics, Space filling by Cu rves, Unequal Circle and Polynomial packing, Computational Geometry and Polygonal packing by Voronoi Tessellation, Four color theorem, Economi c Merit, Computational Economics - GDP estimation from electricity consumption, Jordan Curve Theorem, Weirstrass Theorem, Cellular Automata-SL EUTH Urban Growth Models, Automatic Delineation of Urban Growth Boundaries, Preferential attachment,Random Graphs,Dynamic Graphs,Separate Chai ning Hash tables, Unmanned Aerial Vehicle Navigation-Autonomous Drone Delivery-Drone EVMs-Passenger Drones,Wiring Problem +-------------------------------------------------------------------------------------------------------------------------------------------
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
--------------------------------------------------------------------------------------------------------------------------------------------- 1252. (THEORY and FEATURE) Merit of Large Scale Visuals - GIS Urban Sprawl Analytics - Automatic Delineation of Urban Growth Boundaries from Transportation Network GIS - related to 1151,1174,1196,1226,1233,1234,1237 and all sections on GIS Urban Sprawl Analytics, Merit of Large Scale Visuals, Gridded Population Estimators (SEDAC GPW4 and GHSL convex hull estimators), DBSCAN and Neural Network Urban Extent Learners, Computational Geometry, Intrinsic Merit Rankings, R Ricker Population Estimation, Community detection, Social Network Analytics, Space filling by Curves, Unequal Circle and Polynomial packing, Computational Geometry and Polygonal packing by Voronoi Tessellation, Four color theorem, Economic Merit, Computational Economics - GDP estimation from electricity consumption, Jordan Curve Theorem, Weirstrass Theorem, Cellular Automata-SLEUTH Urban Growth Models, Automatic Delineation of Urban Growth Boundaries, Preferential attachment,Random Graphs,Dynamic Graphs,Separate Chaining Hash tables, Unmanned Aerial Vehicle Navigation-Autonomous Drone Delivery-Drone EVMs-Passenger Drones,Wiring Problem -------------------------------------------------------------------------------------------------------------------------------------------- 1251.1 An algorithm for Automatic Delineation of Urban Growth Boundaries(UGB) from Minimum Spanning Tree of Building Coverage Ratio Segmentation (Meta Research - High Resolution Settlement Layer) and Voronoi-Delaunay Tessellation-Triangulation vis-a-vis Expert Delineation has been studied in 1226.19 - Algorithmic UGB Delineation has been found to be ~93% accurate at its maximum (Frankfurt-Main) versus manual expert delineation. 1251.2 Extracting Road networks from GIS (a crucial prerequisite for predefined Unmanned Aerial Vehicle mission flight plans) preserving topological orientation is of paramount significance in delimiting rural-urban boundaries. 1251.3 Transport Networks are graph structures with added topological embedding defined on a 2D plane connecting urban centres (allowing twists and turns in each edge as against traditional straightline edges of Graph theory) and an optimal layout of transportation network is an unrestricted-twist-and-turn variant of wiring problem in logic gate arrays (optimal space efficient layout of wires connecting logic gates is reduced to optimal space efficient layout of transport network carriageways connecting urban centres) which is NP-Hard for maximum of one 90 degree turn. Problem of unrestricted 90 degree turns between two points approximating a carriageway could be significantly harder(a random walk on a grid). Usual graph algorithms discard layout topology. 1251.4 Inferring a graph structure which preserves topology of transportation network edges from a Transit network GIS is non-trivial and similar to sections 1233,1237 (which warrant a Graph topology from Quadmesh that preserves face orientations),contour polynomials (which correspond to topological orientation of road-rail network) of transportation network are retrieved from segmentation of an example GIS image of Chennai Metropolitan Area (python-src/image_pattern_mining/ImageNet/testlogs/ChennaiMetropolitanAreaTransitNetwork_GoogleMaps_20May2022.jpg - Google Maps augmented by Transport network,Terrain,Live Traffic layers) bypassing necessity for access to Graph structure of Transportation network (e.g Google Roads REST API) 1251.5 Automatic Delineation of Chennai Metropolitan Area by NASA VIIRS NightTime Lights based Urban Extent Learner in earlier sections (1174)depends on luminosity of night lights which at times could be disproportionate to urban area and only a lowerbound of the sprawl. Earlier Google Maps GIS of Transportation network - python-src/image_pattern_mining/ImageNet/testlogs/ChennaiMetropolitanAreaTransitNetwork_GoogleMaps_20May2022.jpg - is a daylight image and vegetation-tree-canopies (NDVI) are conspicuous by green color pixels. Extent of vegetation is inversely proportional to urbanization and degree of transportation network is maximum in urban centres (as multiple carriageways originate from a thriving megapolis). 1251.6 Extent of Urbanization is proportional to dense subgraphs of transportation networks - In earlier GIS imagery, high concentration of transit network aligned to Chennai Metropolitan Area sprawl of approximately 100 kms radius (coinciding with suburban transport network) in North-South-West has been reflected in topological clustering of contour polynomials of transit network plotted in ChennaiMetropolitanAreaTransitNetwork_GoogleMaps_20May2022-contoursegmented.jpg. Extent of a Metropolitan Area Sprawl has been defined earlier (1151) as area catered by suburban transport network or floating population depending on urban centre. 1251.7 Partitioning of contour polynomials of transportation network edges into dense and sparse clusters is a delineator: (*) Dense contour polynomial clusters (subgraphs) of small face areas denote an urban sprawl (*) Sparse contour polynomial clusters (subgraphs) of high face areas might be rural. Voronoi Tessellation-Delaunay Triangulation of earlier image at ChennaiMetropolitanAreaTransitNetwork_GoogleMaps_20May2022-contourlabelled.jpg is too fine-grained and does not show transportation network clustering. 1251.8 Previous set of contour polynomials of transportation network could (*) either be an automatic delineator of Chennai Metropolitan Area Sprawl based on clusters of contour polynomials segregated by their Hausdorff topological distance or Inner product distance (*) or assist a manual expert delineation. 1251.9 ImageGraph_Keras_Theano.py is directly imported instead of "from-import" earlier. Following are logs,facegraph DOT file,contour-Voronoi-Delaunay-watershed segmentation imagery for this commit: testlogs/ChennaiMetropolitanAreaTransitNetwork_GoogleMaps_20May2022-contourlabelled.jpg testlogs/ChennaiMetropolitanAreaTransitNetwork_GoogleMaps_20May2022-contoursegmented.jpg testlogs/ChennaiMetropolitanAreaTransitNetwork_GoogleMaps_20May2022.jpg testlogs/GISUrbanSprawlAnalytics.log.20May2022 testlogs/RemoteSensingGIS/ChennaiMetropolitanAreaTransitNetwork_GoogleMaps_20May2022_ImageNet_Keras_Theano_Segmentation_FaceGraph.dot testlogs/RemoteSensingGIS/ChennaiMetropolitanAreaTransitNetwork_GoogleMaps_20May2022_segmented.jpg
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
NeuronRain Design Section numbers added to all text files
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
Section 1226 updated
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
+--------------------------------------------------------------------------------------------------------------------------------------------- +1240. (THEORY and FEATURE) Merit of Large Scale Visuals - Archaeology - Rebus Decipherment of Indus Unicorn Seals and their validation by + extracting common subgraphs in ImageNet predictions textgraphs (GSpan Frequent Subgraph Mining algorithm) - 13 May 2022 - related to 1203,1232,1233,1236,1237 and all sections on Merit of Large Scale Visuals, Physique Recognition From Quadrilateral mesh isomorphism,Topological Face and Handwriting recognition, Archaeology-Topological Script Recognition, Product Homotopy, Homeomorphism and Topological Deformations, People Analytics, Computational Geometry, Graph Edit Distance and Earth mover distance, Graph Isomorphism (Exact, Approximate and Subgraph),Graph matching,Graph similarity,Image Search,Quadrilateral Mesh Deformation and Isomorphism,Maximum Common Subgraph Problem and Approximate Topological Matching +---------------------------------------------------------------------------------------------------------------------------------------------
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
--------------------------------------------------------------------------------------------------------------------------------------------- 1240. (THEORY and FEATURE) Merit of Large Scale Visuals - Archaeology - Rebus Decipherment of Indus Unicorn Seals and their validation by extracting common subgraphs in ImageNet predictions textgraphs (GSpan Frequent Subgraph Mining algorithm) - 13 May 2022 - related to 1203,1232,1233,1236,1237 and all sections on Merit of Large Scale Visuals, Physique Recognition From Quadrilateral mesh isomorphism,Topological Face and Handwriting recognition, Archaeology-Topological Script Recognition, Product Homotopy, Homeomorphism and Topological Deformations, People Analytics, Computational Geometry, Graph Edit Distance and Earth mover distance, Graph Isomorphism (Exact, Approximate and Subgraph),Graph matching,Graph similarity,Image Search,Quadrilateral Mesh Deformation and Isomorphism,Maximum Common Subgraph Problem and Approximate Topological Matching --------------------------------------------------------------------------------------------------------------------------------------------- 1240.1 New function frequent_subgraphs_of_imagenet_prediction_textgraphs() has been defined to compute GSpan frequent subgraph mining of textgraphs deeplearnt from ImageNet predictions (ResNet50,ResNet50V2) by Recursive Gloss Overlap Dense subgraph classifier. Frequent subgraphs of ImageNet prediction textgraphs learnt from different inscriptions approximately correspond to common scripts-motifs in the inscriptions. Frequent subgraph mining generalizes set intersection of Rebus pictogram textgraphs mentioned in 1238. 1240.2 Earlier textgraph mining function is invoked on 3 archaeological artefacts - Indus Unicorn Seals of differing scripts having some common motifs: image_pattern_mining/ImageNet/testlogs/bagasra-unicorn-seal.jpg, image_pattern_mining/ImageNet/testlogs/unicorn-seal_0.jpg, image_pattern_mining/ImageNet/testlogs/unicorn-seal_1_5.jpg 1240.3 Mined Common Subgraphs (ResNet50 and ResNet50V2) throw some light on intended meaning of the unicorn seals (Disclaimer: it is an automated machine learning decipherment based on accuracy of Rebus Image Search and shouldn't be construed as a full-fledged decipherment claim) - full logs at testlogs/ImageGraph_Keras_Theano.log.13May2022 : ResNet50: Mined Frequent Subgraph Edges: {('brass', 'tube'), ('times', 'trilobite'), ('igniter', 'used'), ('igniter', 'substance'), ('instrument', 'brass'), ('sundial', 'dial'), ('bolo_tie', 'neck'), ('printed', 'book_jacket'), ('parts', 'trilobite'), ('worn', 'bolo_tie'), ('trilobite', 'abundant'), ('brass', 'wind'), ('book', 'book_jacket'), ('jacket', 'book_jacket'), ('brass', 'cup-shaped'), ('sundial', 'indicates'), ('substance', 'igniter'), ('three', 'trilobite'), ('cup-shaped', 'brass'), ('means', 'brass'), ('obelisk', 'pyramidal'), ('trilobite', 'exoskeleton'), ('arthropod', 'trilobite'), ('blown', 'brass'), ('ignite', 'igniter'), ('usually', 'brass'), ('trilobite', 'times'), ('book_jacket', 'information'), ('daylight', 'sundial'), ('brass', 'instrument'), ('tapering', 'obelisk'), ('rectangular', 'obelisk'), ('Paleozoic', 'trilobite'), ('obelisk', 'pillar'), ('neck', 'bolo_tie'), ('variable', 'brass'), ('timepiece', 'sundial'), ('trilobite', 'Paleozoic'), ('base', 'support'), ('book_jacket', 'jacket'), ('ornamental', 'bolo_tie'), ('book_jacket', 'paper'), ('brass', 'usually'), ('obelisk', 'cross'), ('trilobite', 'extinct'), ('bolo_tie', 'necktie'), ('necktie', 'bolo_tie'), ('obelisk', 'stone'), ('cord', 'bolo_tie'), ('obelisk', 'towards'), ('hours', 'sundial'), ('buckle', 'buckles'), ('gnomon', 'sundial'), ('fasten', 'buckle'), ('calibrated', 'sundial'), ('obelisk', 'top'), ('foundation', 'base'), ('stone', 'obelisk'), ('book_jacket', 'promotional'), ('brass', 'blown'), ('clasp', 'bolo_tie'), ('obelisk', 'tapering'), ('cross', 'obelisk'), ('mouthpiece', 'brass'), ('promotional', 'book_jacket'), ('brass', 'funnel-shaped'), ('sundial', 'daylight'), ('brass', 'means'), ('sundial', 'gnomon'), ('bolo_tie', 'clasp'), ('shadow', 'sundial'), ('pyramidal', 'obelisk'), ('divided', 'trilobite'), ('around', 'bolo_tie'), ('casts', 'sundial'), ('fastened', 'bolo_tie'), ('base', 'foundation'), ('extinct', 'trilobite'), ('usually', 'book_jacket'), ('obelisk', 'rectangular'), ('towards', 'obelisk'), ('pillar', 'obelisk'), ('brass', 'brass'), ('book_jacket', 'printed'), ('igniter', 'kindle'), ('bolo_tie', 'cord'), ('book_jacket', 'book'), ('igniter', 'fire'), ('exoskeleton', 'trilobite'), ('dial', 'sundial'), ('sundial', 'casts'), ('igniter', 'ignite'), ('trilobite', 'three'), ('sundial', 'hours'), ('top', 'obelisk'), ('bolo_tie', 'worn'), ('bolo_tie', 'ornamental'), ('brass', 'mouthpiece'), ('fire', 'igniter'), ('trilobite', 'divided'), ('section', 'obelisk'), ('brass', 'length'), ('abundant', 'trilobite'), ('length', 'brass'), ('funnel-shaped', 'brass'), ('trilobite', 'arthropod'), ('sundial', 'timepiece'), ('indicates', 'sundial'), ('information', 'book_jacket'), ('tube', 'brass'), ('wind', 'brass'), ('buckle', 'fasten'), ('paper', 'book_jacket'), ('bolo_tie', 'around'), ('trilobite', 'parts'), ('buckle', 'buckle'), ('sundial', 'shadow'), ('kindle', 'igniter'), ('brass', 'consists'), ('brass', 'variable'), ('sundial', 'calibrated'), ('used', 'igniter'), ('obelisk', 'section'), ('bolo_tie', 'fastened'), ('book_jacket', 'usually'), ('consists', 'brass'), ('support', 'base'), ('buckles', 'buckle')} ResNet50V2: Mined Frequent Subgraph Edges: {('common', 'wood_rabbit'), ('barrel', 'hold'), ('wood_rabbit', 'North'), ('English_foxhound', 'hunt'), ('small', 'ice_lolly'), ('wood_rabbit', 'Ixodes'), ('chambered_nautilus', 'spiral'), ('host', 'wood_rabbit'), ('fox_squirrel', 'arboreal'), ('wood_rabbit', 'underside'), ('pale', 'chambered_nautilus'), ('someone', 'hotdog'), ('ice_lolly', 'ice'), ('scapularis', 'wood_rabbit'), ('slightly', 'English_foxhound'), ('English_foxhound', 'larger'), ('trifle', 'sponge'), ('attention', 'hotdog'), ('stick', 'ice_lolly'), ('fruit', 'trifle'), ('cream', 'ice_lolly'), ('chocolate', 'trifle'), ('hotdog', 'attract'), ('barrel', 'size'), ('wood_rabbit', 'rabbit'), ('trifle', 'pudding'), ('wood_rabbit', 'ticks'), ('hunt', 'English_foxhound'), ('stunts', 'hotdog'), ('Pacific', 'chambered_nautilus'), ('nuts', 'trifle'), ('cream', 'trifle'), ('hotdog', 'performs'), ('chambered_nautilus', 'cephalopod'), ('made', 'trifle'), ('United', 'fox_squirrel'), ('trifle', 'cake'), ('wood_rabbit', 'greyish'), ('wood_rabbit', 'common'), ('larger', 'English_foxhound'), ('wood_rabbit', 'small'), ('fox_squirrel', 'States'), ('wood_rabbit', 'fur'), ('brownish', 'wood_rabbit'), ('trifle', 'jelly'), ('water', 'ice_lolly'), ('cephalopod', 'chambered_nautilus'), ('States', 'fox_squirrel'), ('wood_rabbit', 'white'), ('performs', 'hotdog'), ('English_foxhound', 'foxhounds'), ('trifle', 'decorated'), ('attract', 'hotdog'), ('chambered_nautilus', 'pale'), ('chambered_nautilus', 'partitions'), ('rabbit', 'wood_rabbit'), ('disease', 'wood_rabbit'), ('wood_rabbit', 'tail'), ('decorated', 'trifle'), ('tail', 'wood_rabbit'), ('fox_squirrel', 'large'), ('hotdog', 'dangerous'), ('wood_rabbit', 'Lyme'), ('packs', 'English_foxhound'), ('hotdog', 'attention'), ('trifle', 'layers'), ('exceptionally', 'fox_squirrel'), ('wood_rabbit', 'pacificus'), ('greyish', 'wood_rabbit'), ('Lyme', 'wood_rabbit'), ('trifle', 'cold'), ('white', 'wood_rabbit'), ('English_foxhound', 'originally'), ('chambered_nautilus', 'oceans'), ('chambered_nautilus', 'shell'), ('wooden', 'ice_lolly'), ('hotdog', 'someone'), ('dangerous', 'hotdog'), ('ice', 'ice_lolly'), ('small', 'wood_rabbit'), ('spiral', 'chambered_nautilus'), ('cake', 'trifle'), ('hold', 'barrel'), ('underside', 'wood_rabbit'), ('foxhounds', 'English_foxhound'), ('ice_lolly', 'water'), ('squirrel', 'fox_squirrel'), ('English_foxhound', 'breed'), ('chambered_nautilus', 'pearly'), ('breed', 'English_foxhound'), ('quantity', 'barrel'), ('ice_lolly', 'small'), ('sponge', 'trifle'), ('wood_rabbit', 'host'), ('English', 'English_foxhound'), ('trifle', 'fruit'), ('English_foxhound', 'American'), ('pacificus', 'wood_rabbit'), ('trifle', 'nuts'), ('chambered_nautilus', 'Pacific'), ('fox_squirrel', 'squirrel'), ('oceans', 'chambered_nautilus'), ('wood_rabbit', 'disease'), ('large', 'fox_squirrel'), ('spread', 'trifle'), ('English_foxhound', 'used'), ('ice_lolly', 'cream'), ('pudding', 'trifle'), ('eastern', 'fox_squirrel'), ('barrel', 'quantity'), ('fur', 'wood_rabbit'), ('wood_rabbit', 'brownish'), ('partitions', 'chambered_nautilus'), ('wood_rabbit', 'scapularis'), ('fox_squirrel', 'United'), ('wood_rabbit', 'America'), ('Ixodes', 'wood_rabbit'), ('fox_squirrel', 'eastern'), ('shell', 'chambered_nautilus'), ('English_foxhound', 'packs'), ('ice_lolly', 'wooden'), ('North', 'wood_rabbit'), ('America', 'wood_rabbit'), ('ticks', 'wood_rabbit'), ('layers', 'trifle'), ('size', 'barrel'), ('American', 'English_foxhound'), ('cold', 'trifle'), ('trifle', 'made'), ('chambered_nautilus', 'Indian'), ('pearly', 'chambered_nautilus'), ('trifle', 'spread'), ('ice_lolly', 'stick'), ('English_foxhound', 'slightly'), ('English_foxhound', 'English'), ('fox_squirrel', 'exceptionally'), ('trifle', 'cream'), ('trifle', 'chocolate'), ('Indian', 'chambered_nautilus'), ('arboreal', 'fox_squirrel'), ('used', 'English_foxhound'), ('jelly', 'trifle'), ('originally', 'English_foxhound'), ('hotdog', 'stunts'), ('barrel', 'barrel')} 1240.4 From earlier frequent subgraph mining, few notable common vertices between ImageNet textgraphs of 3 unicorn seals are: ResNet50 - sundial, gnomon, fire, cup-shaped (perhaps hinting at an animal sacrifice fire altar and astronomical configuration during sacrifice guided by gnomon-sundial - script on seal might be a description of sacrifice) ResNet50V2 - foxhound (unicorn bull or buffalo figurine has been predicted as foxhound)
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
Section 1150 updated
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
Section 1150 updated
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
Section 2 and 1234 updated
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
-------------------------------------------------------------------------------------------------------------------------------------------- 1239. (THEORY and FEATURE) Merit of Large Scale Visuals - Astronomy Analytics - DBSCAN Clustering and Segmentation analysis of sky imageries for 2 high magnitude seismic events - 9 May 2022 and 10 May 2022 - related to 3,6,8,15,16,59,62,63,147,148 and all sections on analytics of textually encoded astronomical datasets, Various algorithms for string analytics, climate analytics, weather analytics, N-Body problem, Computational Astrophysics, Computational Geometry --------------------------------------------------------------------------------------------------------------------------------------------
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
-------------------------------------------------------------------------------------------------------------------------------------------- 1239. (THEORY and FEATURE) Merit of Large Scale Visuals - Astronomy Analytics - DBSCAN Clustering and Segmentation analysis of sky imageries for 2 high magnitude seismic events - 9 May 2022 and 10 May 2022 - related to 3,6,8,15,16,59,62,63,147,148 and all sections on analytics of textually encoded astronomical datasets, Various algorithms for string analytics, climate analytics, weather analytics, N-Body problem, Computational Astrophysics, Computational Geometry -------------------------------------------------------------------------------------------------------------------------------------------- 1239.1 NeuronRain has so far been confined to analytics of textual astronomical datasets correlated to terrestial extreme weather events, which are encoded strings of degree alignments (N-body choreography) of celestial bodies on zodiac obtained from an ephemeris client (NASA JPL ephemeris - Maitreya's Dreams Text Client). 1239.2 This commit extends astronomy analytics to sky imagery of celestial bodies sourced from Planetarium software. Astronomical imagery of celestial bodies are scatterplots of distant objects (constellations,planets,meteor showers are usually circles of varying radii spread out on sky) and visual clustering and segmentation algorithms viz., DBSCAN, Watershed segmentation, Chain approximation contour detection could be suited for classifying the point celestial objects into en-masse clusters. Each cluster thus found is a candidate for gravitational influence on earth's tectonics which in a sense generalizes 12 house static zodiac division over 360 degrees of 30 degrees each, to a sky arbitrarily divided by clustering and segmentation algorithms based on density of objects. 1239.3 Celestial objects for two high richter scale seismic events and two of the top 10 high deathtoll disasters ever - Tangshan earthquake of 26 July 1976 and Boxing Day Sumatra earthquake of 26 December 2004 (both coinciding with fullmoon or newmoon - fullmoon or newmoon is often cited as prerequisite for a seismic event possibly exacerbated by a solar or lunar eclipse) - have been plotted in The Sky Live planetarium as observed from respective epicentres and captured in two JPG files (includes both above horizon and below horizon bodies) - SkyLive_SumatraBoxingDayEarthQuake_26December2004_0758.jpg (7:58 am local time) and SkyLive_TangshanEarthQuake_28July1976_0342.jpg (3:42 am local time) 1239.4 Image segmentation in ImageGraph_Keras_Theano.py has been changed to include Delaunay Triangulation and Voronoi Tessellation plots for an image imported from GISUrbanSprawlAnalytics.py. Keras backend has been set to Theano. Return value tuple of image_segmentation() has been updated to return img ndarray and contourcentroids instead of GRANA centroids. 1239.5 Following computation has been performed on 2 seismic events sky imagery (logs,Voronoi and Delaunay imagery,DBSCAN clustering imagery, Neural network clustering imagery,DOT files of the facegraphs): 1239.5.1 Histogram Earth Mover Distance similarity between how sky looked like when two seismic events struck - testlogs/ImageGraph_Keras_Theano.AstronomicalImagery.log.9May2022 - ('EMD Histogram distance similarity between images - testlogs/SkyLive_SumatraBoxingDayEarthQuake_26December2004_0758.jpg and testlogs/SkyLive_ TangshanEarthQuake_28July1976_0342.jpg :', (7.763535976409912, None, array([[1., 0., 0., ..., 0., 0., 0.]...) 1239.5.2 DBSCAN clustering of the two sky imagery - testlogs/SkyLive_SumatraBoxingDayEarthQuake_26December2004_0758_dbscanclustered.jpg, testlogs/SkyLive_TangshanEarthQuake_28July1976_0342_dbscanclustered.jpg - number of DBSCAN clusters differ slightly between two imagery as: number of DBSCAN density clusters: 104719 size of image: 869031 size of dbscandict: 446804 number of DBSCAN density clusters: 105352 size of image: 869031 size of dbscandict: 446813 1239.5.3 Neural network clustering of the two sky imagery - testlogs/SkyLive_SumatraBoxingDayEarthQuake_26December2004_0758_neuralnetworkclustered.jpg, testlogs/SkyLive_TangshanEarthQuake_28July1976_0342_neuralnetworkclustered.jpg 1239.5.4 Watershed segmentation of the two sky imagery - testlogs/RemoteSensingGIS/SkyLive_SumatraBoxingDayEarthQuake_26December2004_0758_segmented.jpg, testlogs/RemoteSensingGIS/SkyLive_TangshanEarthQuake_28July1976_0342_segmented.jpg 1239.5.5 Voronoi tessellation and Delaunay triangulation around centroids of Chain approximation contours of the two sky imagery - testlogs/SkyLive_SumatraBoxingDayEarthQuake_26December2004_0758-contourlabelled.jpg, testlogs/SkyLive_TangshanEarthQuake_28July1976_0342-contourlabelled.jpg 1239.5.6 DOT files of the facegraphs of 2 imagery - testlogs/RemoteSensingGIS/SkyLive_SumatraBoxingDayEarthQuake_26December2004_0758_ImageNet_Keras_Theano_Segmentation_FaceGraph.dot, testlogs/RemoteSensingGIS/SkyLive_TangshanEarthQuake_28July1976_0342_ImageNet_Keras_Theano_Segmentation_FaceGraph.dot 1239.6 Of all the clustering data earlier, Watershed segmentation imagery has computed a remarkable visual demarcation of celestial clustering compared to DBSCAN clustering, Neural network clustering, Voronoi-Delaunay tessellation pivoted around Chain approximation contour centroids (which is too fine-grained mesh). Additionally what catches the attention is the unusually high number of meteor showers in Tangshan imagery though clustering is sparse (7.5 richter), while Sumatra imagery has noticeable clustering of bodies around Milky way and less meteor showers (9.1 richter).
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
Section 1237 updated
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
Section 1217 updated
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
Section 1238 updated
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
---------------------------------------------------------------------------------------------------------------------------------------------- 1238. (THEORY and FEATURE) Merit of Large Scale Visuals - ImageNet predictions from multiple models - Rebus Topological Script Recognition - Dholavira Signboard example deciphered by ResNet50 and ResNet50V2 - 28 April 2022 - relateed to 1236,1237 and all sections on Archaeology-Topological Script Recognition, Face-Physique-Handwriting-Gender-Motif Recognition, Product Homotopy, Unique identification, Homeomorphism and Topological Deformations, Merit of Large Scale Visuals ---------------------------------------------------------------------------------------------------------------------------------------------
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
---------------------------------------------------------------------------------------------------------------------------------------------- 1238. (THEORY and FEATURE) Merit of Large Scale Visuals - ImageNet predictions from multiple models - Rebus Topological Script Recognition - Dholavira Signboard example deciphered by ResNet50 and ResNet50V2 - 28 April 2022 - relateed to 1236,1237 and all sections on Archaeology-Topological Script Recognition, Face-Physique-Handwriting-Gender-Motif Recognition, Product Homotopy, Unique identification, Homeomorphism and Topological Deformations, Merit of Large Scale Visuals --------------------------------------------------------------------------------------------------------------------------------------------- 1238.1 ImageGraph_Keras_Theano.py has been updated to refactor imagenet_imagegraph() for ImageNet predictions from multiple Deep Learning Models instead of hardcoded ResNet50 earlier. 1238.2 Models have been instantiated within a dictionary of modelname-model map and are iterated in a loop and ImageNet predictions for an image are sequentially printed. List of all available Keras Deep Learning models and their top-k accuracy are at https://keras.io/api/applications/ 1238.3 Since Deep Learning models are efficient probabilistic guesses of artifacts in an image based on convolution network layers, top accuracy has been increased to 5 to allow top 5 probabilities. 1238.4 To verify the Rebus principle topological script recognition algorithm mentioned in 1236 and 1237, Dholavira signboard image from https://www.harappa.com/blog/dholavira-and-its-mysterious-sign-board (testlogs/The_Ten_Indus_Scripts_discovered_near_the_northern_gateway_of_the_Dholavira_citadel.jpg) containing 10 pictorial symbols of Indus script has been deciphered by ResNet50 and ResNet50V2 models in imagenet_imagegraph() and the results ensue (full logs: testlogs/ImageGraph_Keras_Theano.log.28April2022 for top-5 and ImageGraph_Keras_Theano.log2.28April2022 for top-1 accuracies). 1238.5 For ResNet50 predictions and their top-5 probabilities are: ('Predictions - ', 'ResNet50', ':', [[('n03857828', 'oscilloscope', 0.18934955), ('n03291819', 'envelope', 0.05882809), ('n02840245', 'binder', 0.029974239), ('n04554684', 'washer', 0.025299871), ('n03041632', 'cleaver', 0.024807546)]]) 1238.6 For ResNet50V2 predictions and their top-5 probabilities are: ('Predictions - ', 'ResNet50V2', ':', [[('n02123597', 'Siamese_cat', 0.9940566), ('n02089973', 'English_foxhound', 0.0059434352), ('n07613480', 'trifle', 1.8534284e-13), ('n02356798', 'fox_squirrel', 4.2353065e-23), ('n04479046', 'trench_coat', 1.3498343e-29)]]) 1238.7 Counterintuitively, ResNet50 version 1 is predicting better than ResNet50 version 2 for top-5 accuracy - Cleaver or a large knife, Washer or a wheel are reasonable matches for symbols in signboard excluding "oscilloscope"- though both are far away from expected decipherment by a hypothetical DP contour polynomial match algorithm mentioned earlier in 1237.8. Extending deep-learning models to train and match arbitrary topological varieties (contour polynomials of images instead of images - a mammoth computational task in itself to create such a contour database) by their polynomial distance similarity might improve decipherment accuracy. 1238.8 Top pageranked vertices in Textgraph of ImageNet-ResNet50 predictions in 1238.5 are: [('oscilloscope', 0.12046130243949144), ('envelope', 0.12046130243949144), ('cleaver', 0.08107992792460737), ('binder', 0.08107992792460737), ('washer', 0.06795280308631269), ('someone', 0.018726084942707544), ('things', 0.018726084942707544), ('washes', 0.018726084942707544), ('living', 0.018726084942707544), ('large', 0.018069728700792813), ('knife', 0.018069728700792813), ('machine', 0.018069728700792813), ('butcher', 0.018069728700792813), ('blade', 0.018069728700792813), ('binds', 0.018069728700792813), ('grain', 0.018069728700792813), ('sheaves', 0.018069728700792813), ('square', 0.018069728700792813), ('cuts', 0.018069728700792813), ('electrical', 0.017085194337920707), ('varying', 0.017085194337920707), ('equipment', 0.017085194337920707), ('usually', 0.017085194337920707), ('rectangular', 0.017085194337920707), ('thin', 0.017085194337920707), ('quantities', 0.017085194337920707), ('images', 0.017085194337920707), ('provides', 0.017085194337920707), ('letter', 0.017085194337920707), ('visual', 0.017085194337920707), ('flat', 0.017085194337920707), ('etc', 0.017085194337920707), ('electronic', 0.017085194337920707), ('package', 0.017085194337920707), ('container', 0.017085194337920707)] rebus-deciphering the signboard to be a natural language phrase for a "butchering or grain shop using equipments (blade,washer,cleaver-knife)" which coincides with visual intuition.
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
Correction to GMSH QuadMesh GEO file
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
Section 1237 updated
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
+--------------------------------------------------------------------------------------------------------------------------------------------- +1237. (THEORY and FEATURE) Merit of Large Scale Visuals - Face recognition by Approximate Topological Match of GMSH Quadmesh images - 25,26 April 2022 - related to 1203,1232,1233,1236 and all sections on Merit of Large Scale Visuals, Physique Recognition From Quadrilateral mesh isomorphism,Topological Face and Handwriting recognition, Archaeology-Topological Script Recognition, Product Homotopy, Unique identification, Homeomorphism and Topological Deformations, Merit of Large Scale Visuals, People Analytics, Computational Geometry, Graph Edit Distance and Earth mover distance, Graph Isomorphism (Exact, Approximate and Subgraph),Graph matching,Graph similarity,Image Search,Quadrilateral Mesh Deformation and Isomorphism,Maximum Common Subgraph Problem and Approximate Topological Matching,Astronomical Datasets +--------------------------------------------------------------------------------------------------------------------------------------------
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
--------------------------------------------------------------------------------------------------------------------------------------------- 1237. (THEORY and FEATURE) Merit of Large Scale Visuals - Face recognition by Approximate Topological Match of GMSH Quadmesh images - 25,26 April 2022 - related to 1203,1232,1233,1236 and all sections on Merit of Large Scale Visuals, Physique Recognition From Quadrilateral mesh isomorphism,Topological Face and Handwriting recognition, Archaeology-Topological Script Recognition, Product Homotopy, Unique identification, Homeomorphism and Topological Deformations, Merit of Large Scale Visuals, People Analytics, Computational Geometry, Graph Edit Distance and Earth mover distance, Graph Isomorphism (Exact, Approximate and Subgraph),Graph matching,Graph similarity,Image Search,Quadrilateral Mesh Deformation and Isomorphism,Maximum Common Subgraph Problem and Approximate Topological Matching,Astronomical Datasets -------------------------------------------------------------------------------------------------------------------------------------------- 1237.1 In this commit, an approximate topological match between two contour segmented GMSH quadmesh images of portraits - python-src/testlogs/ExamplePortrait1.jpg and python-src/image_pattern_mining/ImageNet/testlogs/ExamplePortrait2.jpg - has been computed by directed Hausdorff distance between set of DP polynomials of the two image contours. 1237.2 python-src/image_pattern_mining/ImageNet/TriMeshToQuadMesh.geo has been updated to write trimesh MSH file (TriMeshToQuadMesh.msh) for python-src/testlogs/ExamplePortrait1.jpg by loading GEO file in GMSH FLTK GUI and option "save Mesh" - Mesh file TriMeshToQuadMesh.msh has been overwritten. 1237.3 TriMesh earlier is recombined to QuadMesh by python 3.7 script TriMeshToQuadMesh.py through commandline - python3.7 TriMeshToQuadMesh.py 2>&1 > testlogs/python3.7.TriMeshToQuadMesh.log.25April2022 - and QuadMesh for ExamplePortrait1.jpg is written to ExamplePortrait1_mesh.jpg by Blossom-FullQuad Recombination algorithm (an FLTK GUI radio button option). Network HTML plot of Delaunay Trimesh has been made optional by boolean flag bezier_animation set as False by default in DeepLearning_ConvolutionNetwork_BackPropagation.py face recognition routine. 1237.4 QuadMesh for ExamplePortrait2.jpg is already present in ExamplePortrait2_mesh.jpg (created by 1233 earlier) which was obtained by Blossom Recombination. 1237.5 DeepLearning_ConvolutionNetwork_BackPropagation.py has been updated to add additional parameters approximate_topological_match=True and match_fraction=0.000001 to facegraph_similarity_metrics() which is read within a new clause to compute approximate DP polynomials for some fraction of contours (set to 0.000001 by default because full image Hausdorff match - match_fraction=1.0 - is computationally intensive as number of contours for all quadfaces is in order of thousands or almost equal to pixel area of the image as each quadface is a rectangular aggregate of few pixel squares) for quadfaces in two quadmeshes and find an approximate topological match by directed Hausdorff distance between sets of contours belonging to two portraits. Graph Isomorphism computation is excluded by this option and clause returns after topological match computation. 1237.6 Getting a Graph object from QuadMesh (mentioned in 1233.5) has been obviated by contour polynomials computation for every quadmesh face which indirectly encode the topology of the face. Contour segmented images of the two quadmesh(s) for ExamplePortrait1.jpg and ExamplePortrait2.jpg - python-src/image_pattern_mining/ImageNet/testlogs/ExamplePortrait2_mesh_contoursegmented_25April2022.jpg and python-src/testlogs/ExamplePortrait1_mesh_contoursegmented_25April2022.jpg - reveal how contours closely align with each quadrilateral face thereby achieving a superlative objective of not just extracting graph structure of the mesh but also preserving topological orientation of each quadrilateral algebraically. 1237.7 DP Contour polynomials for every quadrilateral face of a quadmesh found earlier is a basic prerequisite for any approximate topological matching algorithm including Hausdorff distance. For example, approximate topological matching of oriented quadfaces in 1203.6 could be augmented by maximum per quadface common contour DP polynomial match which is a further finetuning of maximum common subgraph matching problem. Matching quadfaces could be determined by inner product space polynomial distances between DP polynomials of the quadfaces - quadfaces of small distances are approximate topological matches and similarity is the percentage of matching quadface contours. 1237.8 Earlier DP polynomial distance measure could be suitably adapted to Rebus Topological Script Recognition in archaeological decipherment by exhaustively searching for least distance approximate topological contour match for an undeciphered script quadmesh image, in pre-defined database of pictogram motif quadmesh contour DP polynomials and concatenate textual description for the motifs into a meaningful sentence - Example: Dholavira signboard in 1236.20 might be Rebus deciphered by a hypothetical contour search-match algorithm as "Wheel-Hammer-Wheel-Pounding Urn-Wheel-Wheel-Sickle" (Indus script for a civilization ranging from time duration of 7000BC - Mehrgarh - is presumed to be right-to-left and logo-syllabic) based on an ImageNet ResNet50 prediction search of the motifs of least contour DP polynomial distance which could be contextually inferred as a natural language sentence "Transport of Processed (Agriculture) Products" after ImageNet prediction Textgraph thesaurus expansion for Wheel(Transport),Hammer,Urn and Sickle(Processing tools mostly used in cultivation) neglecting the repetitive logos and their order. 1237.9 Disclaimer: Earlier Dholavira signboard example is not a decipherment claim but stated herein as only a walkthrough on how an AI Topological script recognition-decipherment algorithm might approximate the meaning of an undeciphered archaological inscription. By Rebus principle, meaning of a pictogram inscription could be inferred independent of natural language underneath and knowledge of syllables for each logo may not be necessary - even if it does, ResNet50 prediction in English for an image motif might have to be translated to corresponding word-syllables of all possible contemporary-medieval languages and a sentence might have to be formed from their concatenation right-to-left or left-to-right to zero-in-on the origin of ancient script - an exponentially daunting computational effort. In the example 1236.2, pictogram "h8" could as well be deciphered though an AI algorithm which searches ResNet50 and concatenates syllable-word predictions for "h" and symbol "8" in any other language and only h+eight=height in English is meaningful. ImageNet-ResNet50 is a standard massive deep learning dataset for supercomputers. 1237.10 A paradoxical implication arises if similarities between Keezhadi Script and Indus Script turn out to be derivatives of one another and are classified into one of contemporary languages (Sanskrit as per decipherment claim 1236.11 or Tamil) by an exhaustive Rebus AI contour polynomial-word-syllable search (which is evident from cursory visual glance): Both Indus and Keezhadi scripts should belong to Tamil or else both should be Sanskrit unless a common unifying script was used in the past by both prima facie hinting at common ancestry of Tamil and Sanskrit (as per a legend, Tamil grammar propounded by Agastya is a simplified form of Panini's Sanskrit grammar with Ka-Cha-Ta-Pa-Yadi vargas removed). 1237.11 Decipherment of archaeological motifs is often inseparable from astronomy and recorded history of eclipses in cuneiforms,copper plates,rock edicts and scripture manuscripts is sizeable requiring dating of the inscriptions by C-14 or AMS (Accelerated Mass Spectrometry) and exhaustive search of astronomical data for planetary alignment in an ephemeris software dating back to few thousand years which is hindered by calendar conflicts(Julian-Gregorian-Sidereal-Solar-Lunar),Software overflow errors,Use of Tropical (Dynamic based on Sun's motion over equinoctical points) and Sidereal (Static) zodiacs. Axiomatically it can be stated,hence,that an accurate estimate of civilization timeline must satisfy all 3 predictions(or post-dictions): Script Decipherment, Astronomy and Carbon dating-AMS must concur.
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
Section 1236 updated
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
Section 1236 updated
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
Section 1236 updated
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
----------------------------------------------------------------------------------------------------------------------------- 1343. (JIRA,THEORY and FEATURE) GIS Urban Sprawl Analytics - Verhulste Population Growth Model - Population Estimator 2 - Chennai Metropolitan Area Sprawl Example - 27 October 2022 - related to 823,1174,1175,1177,1179,1194,1211,1234,1290,1291,1292,1295,1296,1297,1301,1304,1312,1313,1320,1323,1326,1328,1342 and all sections on GIS Urban Sprawl Analytics, Merit of Large Scale Visuals, Gridded Population Estimators (SEDAC GPW4 and GHSL convex hull estimators), DBSCAN and Neural Network Urban Extent Learners, Computational Geometry, Intrinsic Merit Rankings, R Ricker Population Estimation, Community detection, Social Network Analytics, Space filling by Curves, Unequal Circle and Polynomial packing, Computational Geometry and Polygonal packing by Voronoi Tessellation, Four color theorem, Economic Merit, Computational Economics - GDP estimation from Urban sprawl demographics, electricity consumption, Poverty alleviation, MARA (Multiple Agent Resource Allocation) Facegraph walk coloring MAXDNFSAT algorithm, Preferential Attachment, Fixed Parameter Tractability-Kernelization, Automatic Delineation of Urban Sprawls, Cellular Automata-SLEUTH-Polya Urn-Dynamic Facegraph Urban Growth Models,Fame versus Merit ----------------------------------------------------------------------------------------------------------------------------- 1343.1 Function verhulste_ricker_population_growth_model(population_at_t1,intrinsic_growth_rate,carrying_capacity,metropolitan_agglomeration,fromyear,toyear) has been enhanced to include range of years for population projection (fromyear to toyear) as arguments. 1343.2 New loop has been introduced to compute Verhulste Population Estimator YoY (Year on Year) assuming constant annual intrinsic growth rate and carrying capacity during fromyear to toyear which is quite a rigid constraint as carrying capacity might increase over the years by infrastructure development (for example if probability-satisfaction method of 1342.8 is applied) and intrinsic growth rate depends on infrastructure and opportunities. for y in range(1,number_of_years): projected_population_at_t2 = population_at_t1 * math.exp(intrinsic_growth_rate*(1 - population_at_t1/carrying_capacity)) print("verhulste_ricker_population_growth_model(): projected " + metropolitan_agglomeration + " (year " + str(fromyear+y) + ") = ",projected_population_at_t2) population_at_t1 = projected_population_at_t2 1343.3 Two different Chennai Metropolitan Area population projection estimates have been derived for year ranges 2020-2050 and 2019-2050 by Verhulste Estimator based on 2 datasources: 1343.3.1 SEDAC Population Estimator v3 - Chennai Metropolitan Area Population Estimates 2020 - 14730872 - verhulste_ricker_population_growth_model(14730872,2.39/100.0,30000000,"Chennai Metropolitan Area Population",fromyear=2020,toyear=2050) - already computed in 1342 but annual intrinsic growth rate percentage 2.39 has been divided by 100. 1343.3.2 Census of India 2011 statistics - Chennai Metropolitan Area Population Estimates 2011 - 15195379 - verhulste_ricker_population_growth_model(15195379,2.39/100.0,30000000,"Chennai Metropolitan Area Population",fromyear=2019,toyear=2050) 1343.4 Notion of intrinsic population growth rate in urban sprawls has uncanny resemblance to intrinsic fitness in social network analysis (ability of a social profile vertex to attract adjacent vertices or Merit counterpart of Fame) 1343.5 Following are population projections from testlogs/GISUrbanSprawlAnalytics.log.27October2022 for CMA sprawl of 5904 sqkm: =================================================================== SEDAC 2020 Estimates =================================================================== verhulste_ricker_population_growth_model(): projected Chennai Metropolitan Area Population (year 2021) = 14911158.614131337 verhulste_ricker_population_growth_model(): projected Chennai Metropolitan Area Population (year 2022) = 15091483.980312804 ... verhulste_ricker_population_growth_model(): projected Chennai Metropolitan Area Population (year 2049) = 19791465.09768546 =================================================================== Census of India 2011 Estimates =================================================================== verhulste_ricker_population_growth_model(): projected Chennai Metropolitan Area Population (year 2020) = 15375659.645372419 verhulste_ricker_population_growth_model(): projected Chennai Metropolitan Area Population (year 2021) = 15555844.825401036 verhulste_ricker_population_growth_model(): projected Chennai Metropolitan Area Population (year 2022) = 15735882.560905077 ... verhulste_ricker_population_growth_model(): projected Chennai Metropolitan Area Population (year 2049) = 20363506.68372657 1343.6 Earlier code changes resolve 2 JIRA issues : https://krishnairesearch.atlassian.net/browse/NEUR-7 and https://krishnairesearch.atlassian.net/browse/NEUR-8
root
More... over 4 years ago
Section 1236 updated
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
+--------------------------------------------------------------------------------------------------------------------------------------------- +1236. (THEORY and FEATURE) GIS Analytics - an archaeoastronomical and computer vision analysis of MODIS GIS image of Adam's Bridge (Ram Sethu) - related to 2 and all sections on Merit of Large Scale Visuals, Text Restoration, Astronomical Datasets, Archaeology, Drone Urban Extent Imagery, Urban Sprawl Analytics +---------------------------------------------------------------------------------------------------------------------------------------------
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
--------------------------------------------------------------------------------------------------------------------------------------------- 1236. (THEORY and FEATURE) GIS Analytics - an archaeoastronomical and computer vision analysis of MODIS GIS image of Adam's Bridge (Ram Sethu) - related to 2 and all sections on Merit of Large Scale Visuals, Text Restoration, Astronomical Datasets, Archaeology, Drone Urban Extent Imagery, Urban Sprawl Analytics --------------------------------------------------------------------------------------------------------------------------------------------- 1236.1 Application of Machine Learning to archaeology has been recently demonstrated by DeepMind Ithaca which restored 2500 year old Greek manuscript by predicting illegible words damaged and lost in antiquity with high accuracy. 1236.2 On similar lines, devising algorithms for deciphering ancient inscriptions (e.g. Voynich Manuscript, Scripts of Sarasvati-Indus Civilization) by artificial intelligence could be challenging open area of research. 1236.3 In this commit, a GIS image of Palk Straits and Adam's Bridge captured by NASA MODIS instrument in 2018 (250 metre resolution - NASA_MODIS_RamSethu_Image07272018_250m.jpg), which is legendarily attributed to Lord Rama, has been photoshopped and filtered by GIMP and contour segmented. An earlier image of Adam's Bridge by NASA in 2007 was subject matter of debate. 1236.4 Adam's Bridge or Ramsethu is perhaps the only available prima facie archaeological evidence for establishing historicity of Ramayana. Differentiating man-made structures from natural terrains has long been investigated by large scale visual algorithms mostly in the context of Drone urban extent imagery analyzers for classification of settlements (1182.6). 1236.5 Proving if aforementioned MODIS image of Adam's Bridge is natural or manmade might require a sophisticated Ground Penetrating Radar (GPR) Remote Sensing going beyond aerial imagery for analysis of ocean floor and depth - scripture mentions the bridge has been built by woods and elephant-sized sandstones dragged by mechanised contrivances and only a GPR imagery might help in verifying if ocean floor has fossil remnants of wooden structures though studies have revealed heavy presence of sandstones. To compensate for lack of GPR original MODIS image has been enhanced by VFX filters (sharpening,anti-aliasing,contrast) to make graded depth of ocean more pronounced. 1236.6 Another curious aspect of Adam's Bridge is its curvature possibly caused by ocean currents, as a bridge is usually a straightline and scripture verses mention that monkey-men used ropes to ensure straightline alignment of the bridge. 1236.7 Contour segmented image NASA_MODIS_RamSethu_Image07272018_250m_contoursegmented_20April2022.jpg has been color coded and colors might mostly be proportional to their depth (or elevation). Particularly the streak of shoals bridging international waters has been uniformly colored in blue. This is probably because of Watershed algorithm which is precisely based on depth (encoded by pixel color) based segmentation. 1236.8 Following are the affected files (Facegraph DOT,segmented image,contour segmented image and logs). Graph plotting by draw_networkx() has been commented and only DOT files are created because of slowdown: testlogs/ImageGraph_Keras_Theano.log.20April2022 testlogs/NASA_MODIS_RamSethu_Image07272018_250m.jpg testlogs/NASA_MODIS_RamSethu_Image07272018_250m_contoursegmented_20April2022.jpg testlogs/RemoteSensingGIS/NASA_MODIS_RamSethu_Image07272018_250m_ImageNet_Keras_Theano_Segmentation_FaceGraph.dot testlogs/RemoteSensingGIS/NASA_MODIS_RamSethu_Image07272018_250m_segmented.jpg
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
Section 1234 updated
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
Section 1234 updated
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago
Updated BestPractices.txt
Krishna iResearch - NeuronRain - K.Srinivasan
More... over 4 years ago