Date of Award


Publication Type

Master Thesis

Degree Name



Electrical and Computer Engineering


Camera Model Identification, Convolutional Neural Networks, Deep Learning, Digital Image Forensics, Image Manipulation Detection, Machine Learning


Jonathan Wu




Camera model identification (CMI) and image manipulation detection are of paramount importance in image forensics as digitally altered images are becoming increasingly commonplace. In this thesis, we propose a novel convolutional neural network (CNN) architecture for performing these two crucial tasks. Our proposed Remnant Convolutional Neural Network (RemNet) is designed with emphasis given on the preprocessing task considered to be inevitable for removing the scene content that heavily obscures the camera model fingerprints and image manipulation artifacts. Unlike the conventional approaches where fixed filters are used for preprocessing, the proposed remnant blocks, when coupled with a classification block and trained end-to-end, learn to suppress the unnecessary image contents dynamically. This helps the classification block extract more robust images forensics features from the remnant of the image. We also propose a variant of the network titled L2-constrained Remnant Convolutional Neural Network (L2-constrained RemNet), where an L2 loss is applied to the output of the preprocessor block, and categorical crossentropy loss is calculated based on the output of the classification block. The whole network is trained in an end-to-end manner by minimizing the total loss, which is a combination of the L2 loss and the categorical crossentropy loss. The whole network, consisting of a preprocessing block and a shallow classification block, when trained on 18 models from the Dresden database, shows 100% accuracy for 16 camera models with an overall accuracy of 98.15% on test images from unseen devices and scenes, outperforming the state-of-the-art deep CNNs used in CMI. Furthermore, the proposed remnant blocks, when cascaded with the existing deep CNNs, e.g., ResNet, DenseNet, boost their performances by a large margin. The proposed approach proves to be very robust in identifying the source camera models, even if the original images are post-processed. It also achieves an overall accuracy of 95.49% on the IEEE Signal Processing Cup 2018 dataset, which indicates its generalizability. Furthermore, we attain an overall accuracy of 99.68% in image manipulation detection, which implies that it can be used as a general-purpose network for image forensic tasks.