Diagram of an Ear

Ear Recognition Research

University of Ljubljana

Joint work of two laboratories:


  • Evaluation and Analysis of Ear Recognition Models: Performance, Complexity and Resource Requirements

    VGG, SqueezeNet, ResNet50

    Fill in and sign this form and send it to ziga.emersic@fri.uni-lj.si with the subject "AWE Request: The VGG, SqueezeNet, ResNet50 Models".

    	title={Evaluation and analysis of ear recognition models: performance, complexity and resource requirements},
    	author={Emer{\v{s}}i{\v{c}}, {\v{Z}}iga and Meden, Bla{\v{z}} and Peer, Peter and {\v{S}}truc, Vitomir},
    	journal={Neural computing and applications},
  • Convolutional Encoder-Decoder Networks for Pixel-Wise Ear Detection and Segmentation

    PED-CED (modified SegNet architecture)

    Fill in and sign this form and send it to ziga.emersic@fri.uni-lj.si with the subject "AWE Request: PED-CED Model".

    	title={Convolutional encoder--decoder networks for pixel-wise ear detection and segmentation},
    	author={Emer{\v{s}}i{\v{c}}, {\v{Z}}iga and Gabriel, Luka L and {\v{S}}truc, Vitomir and Peer, Peter},
    	journal={IET Biometrics},

Book Chapters

  • Constellation-Based Deep Ear Recognition


    Fill in and sign this form and send it to ziga.emersic@fri.uni-lj.si with the subject "AWE Request: COM-Ear Model".

    	author="{\v{S}}tepec, Dejan
    	and Emer{\v{s}}i{\v{c}}, {\v{Z}}iga
    	and Peer, Peter
    	and {\v{S}}truc, Vitomir",
    	editor="Jiang, Richard
    	and Li, Chang-Tsun
    	and Crookes, Danny
    	and Meng, Weizhi
    	and Rosenberger, Christophe",
    	title="Constellation-Based Deep Ear Recognition",
    	bookTitle="Deep Biometrics",
    	publisher="Springer International Publishing",
    	abstract="This chapter introduces COM-Ear, a deep constellation model for ear recognition. Different from competing solutions, COM-Ear encodes global as well as local characteristics of ear images and generates descriptive ear representations that ensure competitive recognition performance. The model is designed as dual-path convolutional neural network (CNN), where one path processes the input in a holistic manner, and the second captures local images characteristics from image patches sampled from the input image. A novel pooling operation, called patch-relevant-information pooling, is also proposed and integrated into the COM-Ear model. The pooling operation helps to select features from the input patches that are locally important and to focus the attention of the network to image regions that are descriptive and important for representation purposes. The model is trained in an end-to-end manner using a combined cross-entropy and center loss. Extensive experiments on the recently introduced Extended Annotated Web Ears (AWEx).",
  • Deep Ear Recognition Pipeline

    RefineNet (segmentation) and ResNet-152 (classification)

    Fill in and sign this form and send it to ziga.emersic@fri.uni-lj.si with the subject "AWE Request: RefineNet and ResNet-152 Models".

    author="Emer{\v{s}}i{\v{c}}, {\v{Z}}iga
    and Kri{\v{z}}aj, Janez
    and {\v{S}}truc, Vitomir
    and Peer, Peter",
    editor="Hassaballah, Mahmoud
    and Hosny, Khalid M.",
    title="Deep Ear Recognition Pipeline",
    bookTitle="Recent Advances in Computer Vision: Theories and Applications",
    publisher="Springer International Publishing",
    abstract="Ear recognition has seen multiple improvements in recent years and still remains very active today. However, it has been approached from recognition and detection perspective separately. Furthermore, deep-learning-based approachesEmer{\v{s}}i{\v{c}}, {\v{Z}}iga that are popular in other domains have seen limited use in ear recognition and even more so in ear detection. Moreover, to obtain a usableKri{\v{z}}aj, Janez recognition system a unified pipeline{\v{S}}truc, Vitomir is needed. The input in such system should be plain images of subjects and thePeer, Peter output identities based only on ear biometrics. We conduct separate analysis through detection and identification experiments on the challenging dataset and, using the best approaches, present a novel, unified pipeline. The pipeline is based on convolutional neural networks (CNN) and presents, to the best of our knowledge, the first CNN-based ear recognition pipeline. The pipeline incorporates both, the detection of ears on arbitrary images of people, as well as recognition on these segmented ear regions. The experiments show that the presented system is a state-of-the-art system and, thus, a good foundation for future real-word ear recognition systems.",

Competitions (Conferences)

  • The Unconstrained Ear Recognition Challenges 2017 and 2019

    VGG (baseline)

    Fill in and sign this form and send it to ziga.emersic@fri.uni-lj.si with the subject "UERC Request: The Baseline VGG Model".

    	title={The Unconstrained Ear Recognition Challenge 2019},
    	author={Emer{\v{s}}i{\v{c}}, {\v{Z}} and SV, A Kumar and Harish, BS and Gutfeter, W and Khiarak, JN and Pacut, A and Hansley, E and Segundo, M Pamplona and Sarkar, S and Park, HJ and others},
    	booktitle={2019 International Conference on Biometrics (ICB)},
    	title={The unconstrained ear recognition challenge},
    	author={Emer{\v{s}}i{\v{c}}, {\v{Z}}iga and {\v{S}}tepec, Dejan and {\v{S}}truc, Vitomir and Peer, Peter and George, Anjith and Ahmad, Adii and Omar, Elshibani and Boult, Terranee E and Safdaii, Reza and Zhou, Yuxiang and others},
    	booktitle={2017 IEEE international joint conference on biometrics (IJCB)},