{"id":{"repo_id":"usf","oai_identifier":"oai:digitalcommons.usf.edu:etd-9269"},"canonical_url":"https://search.ndltd.org/etd/usf/oai:digitalcommons.usf.edu:etd-9269","repository":{"repo_id":"usf","name":"University of South Florida","base_url":"https://digitalcommons.usf.edu/do/oai/"},"display":{"title":"Detecting Digitally Forged Faces in Online Videos","abstract":"We use Rossler’s FaceForensics dataset of 1004 online videos and their corresponding forged counterparts [1] to investigate the ability to distinguish digitally forged facial images from original images automatically with deep learning. The proposed convolutional neural network is much smaller than the current state-of-the-art solutions. Nevertheless, the network maintains a high level of accuracy (99.6%), all while using the entire FaceForensics dataset and not including any temporal information. We implement majority voting and show the impact on accuracy (99.67%), where only 1 video of 300 is misclassified. We examine why the model misclassified this one video. In terms of tuning the network, we observe how changing hyperparameters affects training time for each epoch and accuracy for training, validation, and testing datasets. There are some challenges involved in obtaining consistent results with deep learning because of the randomization involved in initializing weights. We also replicate Rossler’s XceptionNet [2] experiment for classifying images as originals or forgeries and examine the underlying issues with his research: using a subset of data that is not representative of the full dataset and lack of generalization because of network overfitting when using transfer learning with XceptionNet. Lastly, we explore future work, including forged audio, different network types, and new image datasets.","abstract_html":"We use Rossler’s FaceForensics dataset of 1004 online videos and their corresponding forged counterparts [1] to investigate the ability to distinguish digitally forged facial images from original images automatically with deep learning. The proposed convolutional neural network is much smaller than the current state-of-the-art solutions. Nevertheless, the network maintains a high level of accuracy (99.6%), all while using the entire FaceForensics dataset and not including any temporal information. We implement majority voting and show the impact on accuracy (99.67%), where only 1 video of 300 is misclassified. We examine why the model misclassified this one video. In terms of tuning the network, we observe how changing hyperparameters affects training time for each epoch and accuracy for training, validation, and testing datasets. There are some challenges involved in obtaining consistent results with deep learning because of the randomization involved in initializing weights. We also replicate Rossler’s XceptionNet [2] experiment for classifying images as originals or forgeries and examine the underlying issues with his research: using a subset of data that is not representative of the full dataset and lack of generalization because of network overfitting when using transfer learning with XceptionNet. Lastly, we explore future work, including forged audio, different network types, and new image datasets.","abstract_has_math":false,"creators":["Sambhu, Neilesh"],"institution":"University of South Florida","degree_name":"Master of Science (M.S.)","degree_level":"Thesis","degree_discipline":"Computer Science and Engineering","degree_department":null,"school":null,"contributors":["Shaun Canavan, Ph.D.","Sudeep Sarkar, Ph.D.","Tempestt Neal, Ph.D."],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2019,"date_issued":"2019-10-31T07:00:00Z","date_published":"2019-10-31T07:00:00Z","updated_at":"2026-11-19T08:51:38Z","subjects":["machine learning","deepfake","deep learning","computer vision","convolutional neural network","Accounting","Artificial Intelligence and Robotics"],"languages":[],"rights":[],"rights_urls":[],"identifier_entries":[],"person_identifier_entries":[]},"links":{"outbound_url":"https://digitalcommons.usf.edu/etd/8072","outbound_label":"Repository record","outbound_source":"dc:identifier","record_url":"https://digitalcommons.usf.edu/etd/8072","record_label":"Repository record","record_kind":"record","record_source":"dc:identifier","download_url":null,"download_label":null,"download_source":null,"access_status":null,"embargo_until":null,"access_note":null},"source_record":{"url":"https://digitalcommons.usf.edu/do/oai/?verb=GetRecord&metadataPrefix=oai_etdms&identifier=oai%3Adigitalcommons.usf.edu%3Aetd-9269","prefix":"oai_etdms"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Shaun Canavan, Ph.D.","Sudeep Sarkar, Ph.D.","Tempestt Neal, Ph.D."]},{"key":"dc:creator","label":"Author","values":["Sambhu, Neilesh"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2019-10-31T07:00:00Z"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Computer Science and Engineering"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Thesis","M.S."]},{"key":"thesis:degree_name","label":"Degree Name","values":["Master of Science (M.S.)"]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["University of South Florida"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["machine learning","deepfake","deep learning","computer vision","convolutional neural network","Accounting","Artificial Intelligence and Robotics"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["https://digitalcommons.usf.edu/etd/8072"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["We use Rossler’s FaceForensics dataset of 1004 online videos and their corresponding forged counterparts [1] to investigate the ability to distinguish digitally forged facial images from original images automatically with deep learning. The proposed convolutional neural network is much smaller than the current state-of-the-art solutions. Nevertheless, the network maintains a high level of accuracy (99.6%), all while using the entire FaceForensics dataset and not including any temporal information. We implement majority voting and show the impact on accuracy (99.67%), where only 1 video of 300 is misclassified. We examine why the model misclassified this one video. In terms of tuning the network, we observe how changing hyperparameters affects training time for each epoch and accuracy for training, validation, and testing datasets. There are some challenges involved in obtaining consistent results with deep learning because of the randomization involved in initializing weights. We also replicate Rossler’s XceptionNet [2] experiment for classifying images as originals or forgeries and examine the underlying issues with his research: using a subset of data that is not representative of the full dataset and lack of generalization because of network overfitting when using transfer learning with XceptionNet. Lastly, we explore future work, including forged audio, different network types, and new image datasets."]},{"key":"dc:title","label":"Title","values":["Detecting Digitally Forged Faces in Online Videos"]}]}],"canonical_facts":{"dc:contributor":["Shaun Canavan, Ph.D.","Sudeep Sarkar, Ph.D.","Tempestt Neal, Ph.D."],"dc:creator":["Sambhu, Neilesh"],"dc:date":["2019-10-31T07:00:00Z"],"dc:description":["We use Rossler’s FaceForensics dataset of 1004 online videos and their corresponding forged counterparts [1] to investigate the ability to distinguish digitally forged facial images from original images automatically with deep learning. The proposed convolutional neural network is much smaller than the current state-of-the-art solutions. Nevertheless, the network maintains a high level of accuracy (99.6%), all while using the entire FaceForensics dataset and not including any temporal information. We implement majority voting and show the impact on accuracy (99.67%), where only 1 video of 300 is misclassified. We examine why the model misclassified this one video. In terms of tuning the network, we observe how changing hyperparameters affects training time for each epoch and accuracy for training, validation, and testing datasets. There are some challenges involved in obtaining consistent results with deep learning because of the randomization involved in initializing weights. We also replicate Rossler’s XceptionNet [2] experiment for classifying images as originals or forgeries and examine the underlying issues with his research: using a subset of data that is not representative of the full dataset and lack of generalization because of network overfitting when using transfer learning with XceptionNet. Lastly, we explore future work, including forged audio, different network types, and new image datasets."],"dc:identifier":["https://digitalcommons.usf.edu/etd/8072"],"dc:subject":["machine learning","deepfake","deep learning","computer vision","convolutional neural network","Accounting","Artificial Intelligence and Robotics"],"dc:title":["Detecting Digitally Forged Faces in Online Videos"],"thesis:degree_discipline":["Computer Science and Engineering"],"thesis:degree_level":["Thesis","M.S."],"thesis:degree_name":["Master of Science (M.S.)"],"thesis:institution_name":["University of South Florida"]},"updated_at":"2026-11-19T08:51:38Z"}