| README.md edited online with Bitbucket |
|
More...
|
about 12 years ago
|
| README.md edited online with Bitbucket |
|
More...
|
about 12 years ago
|
| Renamed Project for publishing |
|
More...
|
about 12 years ago
|
| GPLv3 License |
|
More...
|
about 12 years ago
|
| Changed comments in headers to reflect the project name change to "Computer Vision Wordsearch Solver" |
|
More...
|
over 12 years ago
|
| typo |
|
More...
|
over 12 years ago
|
| Renamed project to Computer Vision Wordsearch Solver |
|
More...
|
over 12 years ago
|
| Improved choice of Algorithm Combinations for Full System Evaluation |
|
More...
|
over 12 years ago
|
| Use the SegmentByMeanDarkPixels segmentation algorithm for wordsearch detection when evaluating the full system |
|
More...
|
over 12 years ago
|
| Write Confusion Matrices out in the standard convention of predicted class being the column and actual class being the row |
|
More...
|
over 12 years ago
|
| Forgot to change the EvaluateWordsearchBitmap method to actually remove small rows and cols from the Segmentation |
|
More...
|
over 12 years ago
|
| Just evaluate one thing at a time (otherwise you can expect a long wait) |
|
More...
|
over 12 years ago
|
| Forgot to change the newly added algorithm combination's name |
|
More...
|
over 12 years ago
|
| Evaluate full system (and Segmentation to Solver stages) with the new RemoveSmallRowsAndCols segmentation option |
|
More...
|
over 12 years ago
|
| Evaluate removing small rows and cols from segmentations impact on the wordsearch detection stage |
|
More...
|
over 12 years ago
|
| Evaluate new method for removing erronious rows and cols from segmentations |
|
More...
|
over 12 years ago
|
| Added functionality to Segmentation object to Remove erroneously small rows and columns |
|
More...
|
over 12 years ago
|
| Wordsearch Detection: Only evaluate on wordsearch images containing one wordsearch (not 2 or more) |
|
More...
|
over 12 years ago
|
| Bug Fix: current & previous networks not being evaluated consistently. Both should have been done on the cross-validation data set (as is now the case) |
|
More...
|
over 12 years ago
|
| Store the previous network AFTER checking for a performance dropoff on the cross-validation data set |
|
More...
|
over 12 years ago
|
| Updated the default algorithms used in the Demo GUI |
|
More...
|
over 12 years ago
|
| Missing AForge.Math library in lib dir |
|
More...
|
over 12 years ago
|
| Evaluate the full system using the best Wordsearch Detection Segmentation Method (as of having the results for the new data) |
|
More...
|
over 12 years ago
|
| Evaluation from Segmentation onwards doesn't need the DetectionSegmentationAlgorithm |
|
More...
|
over 12 years ago
|
| Updated Wordsearch Detection system to no longer incorrectly call itself "Wordsearch Recognition" in logs |
|
More...
|
over 12 years ago
|
| Use newly collected data in the evaluation data set |
|
More...
|
over 12 years ago
|
| Evaluate the stage sof the system after wordsearch detection (Segmentation to Solver) |
|
More...
|
over 12 years ago
|
| Only use images that don't contain wordsearches used in Training or Cross-Validation when evaluating the full system |
|
More...
|
over 12 years ago
|
| Store imag epaths relative to the Image Markup Database data directory |
|
More...
|
over 12 years ago
|
| Missed file |
|
More...
|
over 12 years ago
|