LA Tran

I did the same thing and got the same problem with you, bro! This source may only work for converting and testing from darknet weights.

I use this code in Tensorflow 2.2 and Keras 2.2.4 and it works properly. Use this command for saving weights: model.save_weights("{model_name}.ckpt") Then load model using: model.load_weights("{model_name}.ckpt") Instead of: utils.load_weights(model, "/yolov3")...

Yeah, I also figured it out a few days ago. Anyway, thank you for confirming it. Best, LA Tran

Hi Ratan, It may happen when your input images are captured in different light conditions. Make sure that the common regions of each two images are in the same light...

All your images are captured in the same exposure, right. But the left and the right sides of the images seem like not similar. I guess your images are from...

I am not sure about this. But I think you should try some pre-processing methods to make your image more homogeneous. Try with histogram equalization.

Waiting for someone successfully dealing with this thing here.

hi BunnyIcon, did you try with [multi_image_pano.py](https://github.com/tranleanh/image-panorama-stitching/blob/master/multi_image_pano.py), you can see the command for running at my instruction there.

Hi, You can try this [project](https://github.com/tranleanh/yolov4-keras). I have tried it, it works fine. Good luck.

Hi, Thank you. I will consider it and make an update when I have time. Thanks for your concern.