0% Complete
صفحه اصلی
/
سی و یکمین کنفرانس بین المللی مهندسی برق
Connective Reconstruction-based Novelty Detection
نویسندگان :
Seyyed Morteza Hashemi
1
Parvaneh Aliniya
2
Parvin Razzaghi
3
1- دانشگاه تحصیلات تکمیلی علوم پایه زنجان
2- University of Nevada, Reno
3- دانشگاه تحصیلات تکمیلی علوم پایه زنجان
کلمات کلیدی :
Novelty class detection،Out-of-distribution detection،Computer vision،Deep learning
چکیده :
Detection of out-of-distribution samples is one of the critical tasks for real-world applications of computer vision. The advancement of deep learning has enabled us to analyze real-world data which contain unexplained samples, accentuating the need to detect out-of-distribution instances more than before. GAN-based approaches have been widely used to address this problem due to their ability to perform distribution fitting; however, they are accompanied by training instability and mode collapse. We propose a simple yet efficient reconstruction-based method that avoids adding complexities to compensate for the limitations of GAN models while outperforming them. Unlike previous reconstruction-based works that only utilize reconstruction error or generated samples, our proposed method simultaneously incorporates both of them in the detection task. Our model, which we call "Connective Novelty Detection" has two subnetworks, an autoencoder, and a binary classifier. The autoencoder learns the representation of the positive class by reconstructing them. Then, the model creates negative and connected positive examples using real and generated samples. Negative instances are generated via manipulating the real data, so their distribution is close to the positive class to achieve a more accurate boundary for the classifier. To boost the robustness of the detection to reconstruction error, connected positive samples are created by combining the real and generated samples. Finally, the binary classifier is trained using connected positive and negative examples. We demonstrate a considerable improvement in novelty detection over state-of-the-art methods on MNIST and Caltech-256 datasets.
لیست مقالات
لیست مقالات بایگانی شده
Scattering by an array of PEC cylinders in ferrite media using spectral technique
Zahra Bahrami - Asghar Keshtkar - Ayaz Ghorbani
بررسی تأثیر اجرای سازوکارهای بهره وری انرژی بر ظرفیت سنجی اقتصادی سیستم تأمین برق تجدیدپذیر برای ایستگاه های پایه مخابرات موبایل
بهروز عظیمی امینی - وحید محتشمی - حسین ابوترابی زارچی
بکارگیری تکنیک کنترل مقاوم جهت طراحی مسیر حرکت خودرو در مانورهای اضطراری ممانعت از برخورد
محمد امین قماشی - رضا کاظمی
A Cost-Effective Solution for Traffic Sign Recognition and Geographic Localization Using a Monocular Camera
Mohadeseh Atyabi - Fardin Ayar - Mahdi Javanmardi
Robust H∞ Control Design for Variable-Speed Wind Turbines Using Bilinear Matrix Inequalities
Hamidreza Javanmardi - Alireza Hamedi - Mahya Rahimzadeh
DOA estimation of acoustic signals using stacked products of cross-correlations and coherence factor
Mojtaba Amiri - Amir Akhavan - Ahmad Tavakol - Ehsan Rouhani
Evaluation of Different Connectivity Methods for Obsessive Compulsive Disorder Diagnosis
Samandokht Rashidi - Amin Abdipourasl - Fatemeh Jamaloo - Reza Rostami
A Transformer less Quadratic Boost DC-DC Converter with Continuous Input Current and a Few Number of Components, Based on Classical Boost and Cuk Converter Suitable for Renewable Applications
Saeed Mahdizadeh - Reza Sharifi Shahrivar - Hossein Gholizadeh - Ebrahim Afjei
Robust Consensus for Descriptor Multi-agent Systems with Uncertainties in all Matrices
Abolfazl Saadati Moghadam - Ehsan Ranjbar - Amir Abolfazl Suratgar - Hajar Atrianfar
طراحی و ساخت سیستم تصویربرداری SAR دایروی موج میلیمتری
علی آقاکثیری - امیرعلی بنایی کاشانی - علی تاجیک - علیرضا کیایی - هنگامه عزیزی - مهدی عندلیبی - سامان غضنفری - محمد فخارزاده
بیشتر
ثمین همایش، سامانه مدیریت کنفرانس ها و جشنواره ها - نگارش 42.8.0