Hello, everyone, I encountered one problem: to detect whether one guy has a scarf around his neck or not. I have video images. The person in the images are with various ages,with various colors and with various kinds of clothes. Is there any efficient method to do that? Thanks.
Please post a reference picture. You can segment out the 'scarf' picture using various methods - including shape, texture, color etc.
I mean please send a picture that you want to recognize the scarf.
any way, you first segment the image. for segmentation you can use any features such as wavelet, etc. and SVM to segment the image.
To understand that it is scarf, you need other information such as recognizing face.
so, you explain for what problem you want to recognize the scarf?
do you just find it and remove it for face recognition? or do you want it for image retrieval?
CNNs have shown very good performance on this kind of problems, of course if you have enough data to train your network.
you just list several methods and see the results. there are pretrained CNNs which are useful for your aim. Just test them. even with few data, you can fine tune the CNN.
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering
- Signal Processing
- Image Processing