Hello yeephycho,
Thank you for you GitHub project to get an idea about face detection evaluation. I have also used SSD mobilenetV2 architecture with a wider Face dataset. I have trained this model on GPU, accuracy is good but I am only able to detect face up to the 2-meter distance. I don't understand how can I achieve to detect faces at more than 2m distance. Please, can you tell me what changes can I do in the dataset or model or its parameters to achieve higher accuracy and also able to detect faces at a larger distance?
Hello yeephycho, Thank you for you GitHub project to get an idea about face detection evaluation. I have also used SSD mobilenetV2 architecture with a wider Face dataset. I have trained this model on GPU, accuracy is good but I am only able to detect face up to the 2-meter distance. I don't understand how can I achieve to detect faces at more than 2m distance. Please, can you tell me what changes can I do in the dataset or model or its parameters to achieve higher accuracy and also able to detect faces at a larger distance?