Independent Study Defense – Sheikh Bakir

Time: 10am, Aug 10, 2017.
Location: Cramer 221.
Student: Sheikh Bakir.

“REAL TIME OBJECT DETECTION USING DEEP NEURAL NETWORKS”
Real time object detection using computer vision is one very active research area, with the challenge of getting better accuracy. Researchers are utilizing different machine learning approaches to detect more accurately objects in real time. An example is in the automated car driving area, real time objects detection is crucial for safe driving. In my work I applied Convolution Neural Net (CNN) approach for real time objects detection. Although traditional CNN has archived real-time objects detection with accuracy around 90%, I aimed at a higher level of accuracy, close to 100%, most of the times. In order to achieve such higher level of detection accuracy, I used the more advanced and well known U-NET CNN model.

Posted on .