10 class weight

Edmond Yoo

2018-09-16 f3aced12fcc41415851ffb463675930cea8dd086

10 class weight

 .gitignore

@@ -2,7 +2,6 @@
*.dSYM
*.csv
*.out
*.png
*.so
mnist/
caffe/
@@ -29,4 +28,3 @@
cmake-build-debug/
CMakeLists.txt

*.weights

 README.md

@@ -89,6 +89,6 @@

Result? Suprisingly effective.

<img src="https://github.com/hj3yoo/darknet/blob/master/figures/4_detection_result_1.jpg" width="360"> <img src="https://github.com/hj3yoo/darknet/blob/master/figures/4_decision_result_2.jpg" width="360"><img src="https://github.com/hj3yoo/darknet/blob/master/figures/4_detection_result_3.jpg" width="360"> <img src="https://github.com/hj3yoo/darknet/blob/master/figures/4_decision_result_4.png" width="360">
<img src="https://github.com/hj3yoo/darknet/blob/master/figures/4_detection_result_1.jpg" width="360"> <img src="https://github.com/hj3yoo/darknet/blob/master/figures/4_detection_result_2.jpg" width="360"><img src="https://github.com/hj3yoo/darknet/blob/master/figures/4_detection_result_3.jpg" width="360"> <img src="https://github.com/hj3yoo/darknet/blob/master/figures/4_detection_result_4.png" width="360">

They're of course slightly worse than annonymous detection and impractical for any large number of cardbase, but it was an interesting approach.

 figures/4_detection_result_4.png

 weights/10_cards/tiny_yolo_10_39500.weights

Binary files differ