Open Hub
Open Hub

 Settings |  Report Duplicate

0
I Use This!
Very Low Activity

Commits : Listings

Analyzed about 8 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
------------------------------------------------------------------------------------------------------------- 603. (THEORY and FEATURE) ImageGraph creation from an Image - Keras-Theano - related to 587 and 588 - EventNet Tensor Products Algorithm for Merit of Large Scale Visuals - 30 September 2018 ------------------------------------------------------------------------------------------------------------- 1.This commit introduces a primitive but crucial implementation of EventNet Tensor Products Algorithm for Large Scale Visuals which creates a text graph from information gained from processing of a visual (.jpg). 2.Keras is a framework for Computer Vision which supports pre-trained ImageNet datasets (VGG, ResNet) annotated to ImageNet vertices by Convolution Neural Networks.ImageNet presently has 14 million vertices. 3.Keras has support for TensorFlow, Theano, CNTK backends. A python implementation which imports Keras and sets the backend to Theano and predicts objects in an image from ImageNet has been committed to image_pattern_mining/ImageNet/ImageGraph_Keras_Theano.py 4.This script loads an image in keras-theano, converts it to an image array, expands its dimensions and predicts the objects in the image by ResNet50 pre-trained ImageNet dataset. 5.Predictions for objects in image are decoded into array of tuples of the form (id, imagenet_vertex, probability). 6.An example prediction for few images is described in : https://www.pyimagesearch.com/2016/08/10/imagenet-classification-with-python-and-keras/ along with probabilities for objects annotated. 7.These predicted ImageNet vertices are concatenated into a text and passed on to RecursiveGlossOverlap_Classifier.py which creates a Text Graph - ImageGraph - by invoking RecursiveGlossOverlapGraph() on this text. 8.This algorithm essentially maps an Image to a TextGraph. Presently only still images are supported - extending to Video is straightforward by getting all frames in a video and computing ImageGraphs for all frames and Tensor Products of the causally related ImageGraphs. 9.logs for few images (one of which is courtesy: https://www.aazp.in/live-streaming/ - WhiteTiger_1.jpg) have been committed to image_pattern_mining/ImageNet/testlogs/ 10.Graphs printed in logs show fairly reasonable inference of prominent entities in image.
Srinivasan Kannan
More... about 8 years ago
Updated link to BSP paper - http://web.mit.edu/6.976/www/handout/valiant2.pdf
Srinivasan Kannan
More... about 8 years ago
Theory - Alternative to PRAMs in Parallel k-mergesort for Computational Geometric Factorization - BSP model ( BSP implementations: 1. http://www.bsp-worldwide.org/ 2. Apache Giraph - http://giraph.apache.org/ 3. Pregel - https://dl.acm.org/citation.cfm?id=1807184)
Srinivasan Kannan
More... about 8 years ago
Theory and Feature - References updated for Approximate Counting and Distinct Elements
Srinivasan Kannan
More... about 8 years ago
Updated Spark Commandline for GraphFrames package
Srinivasan Kannan
More... about 8 years ago
Updated Requirements.txt
Srinivasan Kannan
More... about 8 years ago
Software Analytics - GraphFrames - GraphX - Parallel Graph Algorithms for CallGraph/CFG/SATURN Graph DOT files
Srinivasan Kannan
More... about 8 years ago
---------------------------------------------------------------------------------------------------------- Software Analytics - GraphFrames - GraphX - Parallel Graph Algorithms for CallGraph/CFG/SATURN Graph DOT files ---------------------------------------------------------------------------------------------------------- 1.NeuronRain till now has a native implementation of parallelized graph processing (e.g definition graph growth implementation in InterviewAlgorithm/ for texts) and sequential NetworkX Graph objects 2.Spark has parallel graph library GraphX and a separate repository GraphFrames which internally wraps GraphX (http://graphframes.github.io/api/python/graphframes.html) 3.This commit introduces an option to choose GraphFrames in analyzing Software Analytics CallGraph/CFG/SATURN DOT files in CyclomaticComplexitySparkMapReducer.py and invokes few parallel algorithm functions for indegrees, strongly connected components, triangle counting. 4.Number of strongly connected components, degrees, triangles in CallGraph/CFG/SATURN graphs are cyclomatic measures of code complexity (both static and dynamic) 5.logs for GraphFrames is committed to testlogs/CyclomaticComplexitySparkMapReducer.log.GraphXFrames.27September2018 6.GraphFrames can be applied not just to Program Analysis DOT files but to any other source (TextGraph, Social Networks etc.,)
Srinivasan Kannan
More... about 8 years ago
Theory - NC-PRAM equivalence, Parallel Point in Planar Subdivision - References
Srinivasan Kannan
More... about 8 years ago
Theory - Algebraic Varieties over Finite Fields and Computational Geometric Factorization
Srinivasan Kannan
More... about 8 years ago
Theory - Algebraic Geometry and Computational Geometric Factorization
Srinivasan Kannan
More... about 8 years ago
Feature - Streaming Algorithms Implementations - Approximate Counting and Distinct Elements
Srinivasan Kannan
More... about 8 years ago
------------------------------------------------------------------------------------------------------------- Streaming Abstract Generator - Absolute paths to Relative paths ------------------------------------------------------------------------------------------------------------- File Datasource paths have been changed from absolute to relative paths
Srinivasan Kannan
More... about 8 years ago
-------------------------------------------------------------------------------------------------------------- Streaming Algorithms - Approximate Counting and Distinct Elements -------------------------------------------------------------------------------------------------------------- 1. Two new python implementations for Streaming Approximate Counting and Distinct Elements in a random subset have been committed to repository 2. These are implementations of two algorithms from JRF presentation notes (on 28-29/10/2010) for course - Topics in Datamining - during August-December 2010, Chennai Mathematical Institute, Chennai. 3. They have been implemented because of their importance for processing Streaming Datasets e.g Spark Streaming
Srinivasan Kannan
More... about 8 years ago
Theory - Parallel Planar Point Location - Factorization in NC - Bridge Separator Tree
Srinivasan Kannan
More... about 8 years ago
Theory - Factorization and Planar Point Location - Persistent Binary Search Trees
Srinivasan Kannan
More... about 8 years ago
Theory - Wavelet Tree and Computational Geometric Factorization - Orthogonal Range Reporting
Srinivasan Kannan
More... about 8 years ago
Theory - Random Restrictions, Switching Lemma, Least Squares SAT Solver, Randomized Rounding
Srinivasan Kannan
More... about 8 years ago
Theory - Hastad Switching Lemma, Least Squares SAT Solver, Random Restrictions - Switching Lemma Intuition
Srinivasan Kannan
More... about 8 years ago
Theory - Hastad Switching Lemma, Least Squares SAT Solver, Random Restrictions
Srinivasan Kannan
More... about 8 years ago
Theory - Hastad Switching Lemma, Least Squares SAT Solver, Random Restrictions
Srinivasan Kannan
More... about 8 years ago
Theory - Wavelet Tree - Tile Summation - Update and Correction
Srinivasan Kannan
More... about 8 years ago
Theory - Program Analysis - Expected execution times of Program Control Statements
Srinivasan Kannan
More... about 8 years ago
Theory - Wavelet Tree - Tile Summation Factorization - Correction
Srinivasan Kannan
More... about 8 years ago
Theory - Wavelet Tree - M-Structures
Srinivasan Kannan
More... about 8 years ago
Updated Requirements.txt
Srinivasan Kannan
More... about 8 years ago
Theory - Sequential Factorization Optimization Error Probability
Srinivasan Kannan
More... about 8 years ago
Updated Requirements.txt
Srinivasan Kannan
More... about 8 years ago
Theory - Sequential Factorization Optimization Error Probability
Srinivasan Kannan
More... about 8 years ago
Theory - Wavelet Tree - Computational Geometry - Range Report
Srinivasan Kannan
More... about 8 years ago