Title :
Inpainting for Big Data
Author :
Bindu, A. ; Kumar, C. N. Ravi
Author_Institution :
Dept. of Comput. Sci. & Eng, Sri Jayachamarajendra Coll. of Eng., Mysore, India
Abstract :
In the current day World with ever rising crime rate, it becomes evident to keep a watch on the secure authentication and retrieval of Human identities. Face Biometrics is a dedicated and one amongst the most popular arena which relies on the facial (biological) traits for unique identity retrieval. Real Time Applications of efficient identity retrieval using face biometric are posed by many hindrances in the form of occlusions/distortions which subsequently result in the drastic reduction of the system response time with the increase in database size. Efficient and error free Real Time face recognition demands the backup of a mammoth database with due considerations to the degrees of occlusion, age variations, distortions, background variations, lighting conditions, moods etc., which pose a major setback to maintain a balance between the Real Time system efficiency, its response time and optimum database size! This lacuna is addressed in the current Novel work of Research to tap the intra facial feature logical relationships to evolve the occluded/distorted facial boundary and features with precision and perfection and also cut short the size of the Real Time Identity Retrieval system database to a great extent!
Keywords :
Big Data; biometrics (access control); face recognition; image reconstruction; image retrieval; very large databases; Big Data; distortions; face biometrics; facial boundary; facial features; inpainting; intrafacial feature logical relationships; mammoth database; occlusions; optimum database size; real time face recognition; real time identity retrieval system database; response time; Algorithm design and analysis; Databases; Face; Facial features; Feature extraction; Image reconstruction; Real-time systems; Big Data; Identity Retrieval; Inpainting; Occlusion; Real Time Face Recognition;
Conference_Titel :
Signal and Image Processing (ICSIP), 2014 Fifth International Conference on
Conference_Location :
Jeju Island
DOI :
10.1109/ICSIP.2014.14