Age estimation and illicit image detection using a stochastic vision model
- Authors: Islam, Mofakharul
- Date: 2013
- Type: Text , Thesis , PhD
- Full Text: false
- Description: The main objective of this research is to investigate and implement a robust approach with a view to provide the Law Enforcement Agencies (LEAs) with a dedicated forensic tool in future for inspecting confiscated PCs from the suspected paedophile to detect pedophilic images automatically and prevent children viewing pornographic and age-inappropriate images at their home and school and adults at their workplace while they are on the Internet. To achieve this goal, we use a novel face descriptor to differentiate child face from adult face based on categorical age specific contextual cues that are based on new knowledge in terms of features or contexts representatives of child and adult face. Given that the craniofacial cues contain enough structural information on visual cues on human face encoded in the form of high level features we can categorize age into adult and children in tandem with low level features. Finally, we will present a novel stochastic vision model based on Markov Random Fields (MRF) prior, which learned the pornographic contextual constraints from the training pornographic images and eventually introduce knowledge on pornography into our proposed stochastic classifier allowing classification of images into pornographic or benign.
- Description: Doctor of Philosophy
Unsupervised color image segmentation using Markov Random Fields Model
- Authors: Islam, Mofakharul
- Date: 2008
- Type: Text , Thesis , Masters
- Full Text:
- Description: We propose a novel approach to investigate and implement unsupervised segmentation of color images particularly natural color images. The aim is to devise a robust unsu- pervised segmentation approach that can segment a color textured image accurately. Here, the color and texture information of each individual pixel along with the pixel's spatial relationship within its neighborhood have been considered for producing precise segmentation of color images. Precise segmentation of images has tremendous potential in various application domains like bioinformatics, forensics, security and surveillance, the mining and material industry and medical imaging where subtle information related to color and texture is required to analyze an image accurately. We intend to implement a robust unsupervised segmentation approach for color im- ages using a newly developed multidimensional spatially variant ¯nite mixture model (MSVFMM) using a Markov Random Fields (MRF) model for improving the over- all accuracy in segmentation and Haar wavelet transform for increasing the texture sensitivity of the proposed approach. [...]
- Description: Master of Computing