Wenzheng Zeng

68-->70 pretrained model for landmark detection on synthetics data

1

Thanks for your great work. I notice that the synthetics dataset that your train the face alignment model has 70 landmark but the ResNet50d model only outputs 68 of them,...

用模拟器运行时报错:Installation failed due to: ''cmd package install-create -r -t -S 40043112' returns error 'Unknown failure: Can't find service: package''

Thanks for your great work. I am not familiar with the youtubevis API for bbox evaluation and want to learn about the evaluation procedure base on the provided trainset annotation....

Thanks for your great work. I just can not figure out where the raw videos are corresponding to the annotations. Can you give a more detailed instruction on how to...

The label files record each action duration with frame level annotation, Is the start frame index 0 or 1?

Questions about the code for JHMDB

4

Thanks for the great work. I have read the code for JHMDB and have some questions: (1) The performance of [email protected] is just 0.72, much lower than the 82.3 that...