Forensic Face Sketch Matching in Criminal Video Database using deep learning techniques
- Version
- Download 9
- File Size 605.76 KB
- File Count 1
- Create Date 22 March 2026
- Last Updated 22 March 2026
Forensic Face Sketch Matching in Criminal Video Database using deep learning techniques
N.D.Gokul Raj
Dept of Computer Science and
Engineering
Dr.M.G.R.Educational and
Research Institute
Chennai,India
gokulraje8@gmail.com
Dr.V.Sai Shanmuga Raja
Dept of Computer Science and
Engineering
Dr.M.G.R.Educational and
Research Institute
Chennai,India
saishanmugaraja.cse@drmgrd
u.ac.in
N.Logesh
Dept of Computer Science and
Engineering
line 3: name of organization
(of Affiliation)
line 4: City, Country
line 5: email address or
ORCID
Dr.M.Sujitha
Dept of Computer Science and
Engineering
Dr.M.G.R.Educational and
Research Institute
Chennai,India
sujitha.ece@drmgrdu.ac.in
M.Tamilmani
Dept of Computer Science and
Engineering
Dr.M.G.R.Educational and
Research Institute
Chennai,India
virattamil@gmail.com
Abstract—In forensic science, it is seen that hand-drawn face sketches are still very limited and time consuming when it comes to using them with the latest technologies used for recognition and identification of criminals.The Forensic face sketches are commonly used when photographic or video evidence is unavailable.Manual matching of these sketches with large scale criminal video database is time- onsuming.So,This paper presents a deep learning-based approach for matching forensic face sketches with faces extracted from survelliance videos. This system as used convolution neuralnetwork and tranfer learning techniques are employed to extract discriminative facial features from both skectch and video frames. This continues with similarity matching is then performed to identify potential suspects. This proposed system reduces human effort and improves identification efficiency and automatically match the drawn composite face sketch with the police database much faster and efficiently using deep learning.video databases are time-consuming and error-prone, whichcan significantly delay criminal investigations.
Keywords— Forensic Sketch , Face Recognition, Deep learning, CNN, Transfer Learning,Video Surveillance.
Download