So, as the title indicates, this is a interest group about computer vision and signal/image processing research community which is limited to the sensor fusion domain in ADAS.

In case ADAS or AV (Autonomous Vehicles), the sensor includes camera, LiDAR, RADAR, Ultrasonic. Each sensor will have different kind of output. These output will differ by units, bandwidth and its properties. Sensor fusion is method which uses a set of algorithms to correlate obtained from the above sensor sources to produce a enhanced single data source which represents the process environment.

I'm very much interested in sharing more about the traditional heuristic algorithms rather than moving to a complete deep learning based approach. I love traditional algorithms than this modern deep learning approaches. Even though the deep learning has gained so much attention these days with the fast growing Gen AI and stuffs, in case computer vision, I believe learning traditional CV methods would encourage us excel in heuristics which is necessary for the preprocessing and postprocessing stages which includes cleaning and annotation.

I'm also having interest in preparing dataset for the above usecases. As an initial step, data annotation can be done for the images. You know what I love photography. I'm practicing photography for past 6 years but not a satisfactory level talent in that. I just want to blend my hobby and professional skillset togather. We shall have the updates regarding that soon.

And, lets start the discussion...

PS: The image embeded above in Marla Singer (A fictional character in Fight Club movie.) which is not a copyright material drawn by Supriya Arunachalam. I love fight club and its philosophy a lot. So, just included it in my homepage.

Regards,

Revanth Arunachalam