Abstract
The system is based on JPEG image encoding and decoding. The frame skipping algorithm is used to quickly obtain the pictures collected by the camera in r eal time. Through the neural network code, the collected low illumination pictures are enhanced. Through the bottom assembly language, the collected original image is compressed into JPEG and saved. Then the JPEG image is read and decompressed into the original image. The underlying assembly statements are vectorized through ASIMD instructions and accelerated in parallel through multithreading. Finally, the number of video frames displayed is equal to the number of video frames collected by the camera. The acceleration effect is very good.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
