some discuss

[zbt1356660588]

Dear Professor,
I'm sorry to disturb you. I am currently studying deep learning, especially object detection and posture recognition. The content we learned in class is related to the identification of criminal behavior. My idea is to recognize behavior through 3D convolution. My approach is to divide the criminal stage into two parts and create two datasets respectively: one for suspicious behavior before a crime, and the other for successful criminal behavior. These two behaviors only last about 3-5 seconds. Then, by preprocessing the video data of these different categories into 16-frame images, along with normal behavior videos, there are three categories in total for training. Finally, a three-classification behavior division will be achieved to predict crime and issue an alarm. However, I'm very interested in your pre-processing part. Did you use OpenCV? Can I discuss it with you in more depth? Sorry for taking up your precious time. Thank you again.
Sincerely
Email: zbt1356660588@gmail.com

[kaledhoshme123]

Dear Professor, I'm sorry to disturb you. I am currently studying deep learning, especially object detection and posture recognition. The content we learned in class is related to the identification of criminal behavior. My idea is to recognize behavior through 3D convolution. My approach is to divide the criminal stage into two parts and create two datasets respectively: one for suspicious behavior before a crime, and the other for successful criminal behavior. These two behaviors only last about 3-5 seconds. Then, by preprocessing the video data of these different categories into 16-frame images, along with normal behavior videos, there are three categories in total for training. Finally, a three-classification behavior division will be achieved to predict crime and issue an alarm. However, I'm very interested in your pre-processing part. Did you use OpenCV? Can I discuss it with you in more depth? Sorry for taking up your precious time. Thank you again. Sincerely Email: zbt1356660588@gmail.com

Welcome
Yes, you need to use OpenCV as a default in many of the video pre-processing methodologies included in the dataset.
I'm available to help with pre-processing steps and more.
I wish you success.