OpenCampus

this is programs for OpenCampus

fast-style-transfer

style transfer demo.

This is based on https://github.com/lengstrom/fast-style-transfer

Firstly, you need to clone from above repository. Then, files in "fast-style-transfer" of this repository are moved to above original repository.

• evaluate_webcam.py ... This is realtime demo using webcamera.

python evaluate_webcam.py --checkpoint <ckpt-model-path>

• demo_webcam.sh ... The usage of evaluate_webcam.py
• make_outputImage.py ... This makes image containing content.jpg, style.jpg, generated-image.py . Each images resized to fixed size.

python make_outputImage.py \
    --content $content_image_file_name \
    --create $created_image_file_name_by_evaluate.jpg \
    --style $style_image_file_name \
    --out $output_file_name

• make_outputImage2.py ... same as make_outputImage.py The difference is that this keeps original image's scale ratio.

• run_all.sh ... Usages of evaluate.py & make_output_image.py

sh run_all.sh

• run_train.sh ... Usage for training model

sh run_train.sh

keras-yolo3

YOLO-v3 implemented with Keras.

This is based on https://github.com/qqwweee/keras-yolo3

Firstly, you need to clone from above repository. Then, files in "keras-yolo3" of this repository are moved to above original repository.

• yolo_webcam.py ... Realtime object detection with YOLO-v3 implemented with Keras.

python yolo_webcam.py

• yolo_tiny_webcam.py ... Realtime objected detection with Tiny-YOLO-v3.

python yolo_tiny_webcam.py

tf-openpose

openpose implemented with tensorflow.

This is based on https://github.com/ildoonet/tf-pose-estimation

Firstly, you need to clone from above repository. Then, replace estimator.py contained in this repository with original tf-pose/estimator.py.

Demo are same as above repository.

python run_webcam.py