Presentations

Automotive Ethernet and Image Processing

The project uses the standard IEEE 802.3 Ethernet to transfer the video data from a camera node to a hand-held device where  image processing is done and processed video is displayed. We made camera node using a Beagle Bone Black which has Soc AM3358/9 at 1GHz with fast Ethernet Port  and an USB camera. A wireless local area network (WLAN) access point is created using a router, here connection between access point and camera node is established via Category 5 cable and connection between access point and Hand-held Device is achieved using IEEE 802.11 standards. We considered camera node as client and hand-held device as server and data transmission is achieved using UDP(User Datagram Protocol). OpenCV is used on the Hand-held Device for image processing. HOG is used for Pedestrian Detection.

What BERTs and GPTs know about your brand? Probing contextual language models for affect associations 

NAACL_Poster_2___savita.pdf

Predicting Sentiments in Image Advertisements Using Semantic Relations Among Sentiment Labels

The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops. June, 2020.