Edmond Yoo
2018-09-07 fc9a45ac615c3be25eca501ca6963a36c317447e
day1 update
4 files modified
5 files added
20 ■■■■ changed files
Makefile 4 ●●●● patch | view | raw | blame | history
README.md 13 ●●●●● patch | view | raw | blame | history
anchors.txt 1 ●●●● patch | view | raw | blame | history
cfg/tiny_yolo.cfg 2 ●●● patch | view | raw | blame | history
darknet patch | view | raw | blame | history
figures/1_augmented_set_example_1.jpg patch | view | raw | blame | history
figures/1_augmented_set_example_2.jpg patch | view | raw | blame | history
figures/1_augmented_set_example_3.jpg patch | view | raw | blame | history
figures/1_augmented_set_example_4.jpg patch | view | raw | blame | history
Makefile
@@ -1,7 +1,7 @@
GPU=1
CUDNN=0
CUDNN=1
CUDNN_HALF=0
OPENCV=0
OPENCV=1
AVX=0
OPENMP=0
LIBSO=0
README.md
@@ -26,4 +26,15 @@
<img src="https://github.com/hj3yoo/darknet/blob/master/figures/0_detection_result_5.jpg" width="360"> <img src="https://github.com/hj3yoo/darknet/blob/master/figures/0_detection_result_6.jpg" width="360"> <img src="https://github.com/hj3yoo/darknet/blob/master/figures/0_detection_result_7.jpg" width="360">
The second and third problems should easily be solved by further augmenting the dataset with random lighting and image skew. I'll have to think more about the first problem, though.
The second and third problems should easily be solved by further augmenting the dataset with random lighting and image skew. I'll have to think more about the first problem, though.
## Day 1
-----------------------
Added several image augmentation techniques to apply to the training set: noise, dropout, light variation, and glaring:
<img src="https://github.com/hj3yoo/darknet/blob/master/figures/1_augmented_set_example_1.jpg" width="360"> <img src="https://github.com/hj3yoo/darknet/blob/master/figures/1_augmented_set_example_2.jpg" width="360"> <img src="https://github.com/hj3yoo/darknet/blob/master/figures/1_augmented_set_example_3.jpg" width="360"> <img src="https://github.com/hj3yoo/darknet/blob/master/figures/1_augmented_set_example_4.jpg" width="360">
Currently trying to generate enough images to start model training. Hopefully this helps.
Recompiled darknet with OpenCV and CUDNN installed, and recalculated anchors.
anchors.txt
New file
@@ -0,0 +1 @@
118.3429,137.0897, 95.8160,181.9724, 140.4955,166.7423, 112.7262,220.6808, 129.2741,198.9876, 159.0679,197.4912, 138.1861,243.0256, 167.4683,229.0091, 165.0264,255.0887
cfg/tiny_yolo.cfg
@@ -115,7 +115,7 @@
activation=linear
[region]
anchors = 0.738768,0.874946,  2.42204,2.65704,  4.30971,7.04493,  10.246,4.59428,  12.6868,11.8741
anchors = 118.3429,137.0897, 95.8160,181.9724, 140.4955,166.7423, 112.7262,220.6808, 129.2741,198.9876, 159.0679,197.4912, 138.1861,243.0256, 167.4683,229.0091, 165.0264,255.0887
bias_match=1
classes=1
coords=4
darknet
Binary files differ
figures/1_augmented_set_example_1.jpg
figures/1_augmented_set_example_2.jpg
figures/1_augmented_set_example_3.jpg
figures/1_augmented_set_example_4.jpg