Deep Low-Shot Learning for Biological Image Classification and Visualization From Limited Training Samples

被引:4
作者
Cai, Lei [1 ]
Wang, Zhengyang [2 ]
Kulathinal, Rob [3 ]
Kumar, Sudhir [3 ]
Ji, Shuiwang [2 ]
机构
[1] Washington State Univ, Sch Elect Engn & Comp Sci, Pullman, WA 99164 USA
[2] Texas A&M Univ, Dept Comp Sci & Engn, College Stn, TX 77843 USA
[3] Temple Univ, Dept Biol, Philadelphia, PA 19122 USA
基金
美国国家科学基金会;
关键词
Training; Task analysis; Feature extraction; Biological system modeling; Deep learning; Annotations; Biological information theory; Biological image classification; deep two-step low-shot learning; Drosophila in situ hybridization (ISH) images; limited training samples; model interpretation and visualization; GENE-EXPRESSION PATTERNS; AUTOMATED ANNOTATION; DROSOPHILA;
D O I
10.1109/TNNLS.2021.3106831
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Predictive modeling is useful but very challenging in biological image analysis due to the high cost of obtaining and labeling training data. For example, in the study of gene interaction and regulation in Drosophila embryogenesis, the analysis is most biologically meaningful when in situ hybridization (ISH) gene expression pattern images from the same developmental stage are compared. However, labeling training data with precise stages is very time-consuming even for developmental biologists. Thus, a critical challenge is how to build accurate computational models for precise developmental stage classification from limited training samples. In addition, identification and visualization of developmental landmarks are required to enable biologists to interpret prediction results and calibrate models. To address these challenges, we propose a deep two-step low-shot learning framework to accurately classify ISH images using limited training images. Specifically, to enable accurate model training on limited training samples, we formulate the task as a deep low-shot learning problem and develop a novel two-step learning approach, including data-level learning and feature-level learning. We use a deep residual network as our base model and achieve improved performance in the precise stage prediction task of ISH images. Furthermore, the deep model can be interpreted by computing saliency maps, which consists of pixel-wise contributions of an image to its prediction result. In our task, saliency maps are used to assist the identification and visualization of developmental landmarks. Our experimental results show that the proposed model can not only make accurate predictions but also yield biologically meaningful interpretations. We anticipate our methods to be easily generalizable to other biological image classification tasks with small training datasets. Our open-source code is available at https://github.com/divelab/lsl-fly.
引用
收藏
页码:2528 / 2538
页数:11
相关论文
共 44 条
  • [1] [Anonymous], 2015, arXiv: Learning
  • [2] Joint stage recognition and anatomical annotation of drosophila gene expression patterns
    Cai, Xiao
    Wang, Hua
    Huang, Heng
    Ding, Chris
    [J]. BIOINFORMATICS, 2012, 28 (12) : I16 - I24
  • [3] Current challenges in open-source bioimage informatics
    Cardona, Albert
    Tomancak, Pavel
    [J]. NATURE METHODS, 2012, 9 (07) : 661 - 665
  • [4] TWO-DIMENSIONAL SPECTRAL-ANALYSIS OF CORTICAL RECEPTIVE-FIELD PROFILES
    DAUGMAN, JG
    [J]. VISION RESEARCH, 1980, 20 (10) : 847 - 856
  • [5] RELATIONS BETWEEN THE STATISTICS OF NATURAL IMAGES AND THE RESPONSE PROPERTIES OF CORTICAL-CELLS
    FIELD, DJ
    [J]. JOURNAL OF THE OPTICAL SOCIETY OF AMERICA A-OPTICS IMAGE SCIENCE AND VISION, 1987, 4 (12): : 2379 - 2394
  • [6] Finn C, 2017, PR MACH LEARN RES, V70
  • [7] Systematic image-driven analysis of the spatial Drosophila embryonic expression landscape
    Frise, Erwin
    Hammonds, Ann S.
    Celniker, Susan E.
    [J]. MOLECULAR SYSTEMS BIOLOGY, 2010, 6
  • [8] Gordon J., 2018, P NEURIPS, P1
  • [9] Hartenstein V., 1993, Atlas of Drosophila Development
  • [10] Hartenstein V., 1995, ATLAS DROSOPHILA DEV