<p>Individual identification of livestock is a key instrument for record keeping, registration, enhancing biosecurity, enabling traceability and performing insurance. In this paper, individual cattle were identified using eye images through iris template matching. The eye images of individual adult crossbred cattle were captured using Nikon (D 5300) model DSLR camera equipped with a macro lens. The images with maximum iris coverage and no glare were selected for template generation. The irises of selected images were segmented and normalized. The normalized segmented images were sent to 1D Log Gabor wavelets for template generation through encoding. All generated templates were stored in template database. One template from the stored templates of each individual cattle was selected as the query template and matched with remaining templates of that individual. The findings revealed with a clear discrimination between individuals, with intra-class matching percentages consistently above 52% and self-matching reached 100% while inter-class matching values remained below 52%.The result obtained in this paper revealed that the individual cattle can be successfully identified using iris. The technology presented in this paper may be used as a tool for iris image based individual cattle identification.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Iris Image-Based Biometric Recognition of Individual Cattle

  • Kandarpa Boruah,
  • Satyendra Nath Mandal,
  • Nilotpal Ghosh

摘要

Individual identification of livestock is a key instrument for record keeping, registration, enhancing biosecurity, enabling traceability and performing insurance. In this paper, individual cattle were identified using eye images through iris template matching. The eye images of individual adult crossbred cattle were captured using Nikon (D 5300) model DSLR camera equipped with a macro lens. The images with maximum iris coverage and no glare were selected for template generation. The irises of selected images were segmented and normalized. The normalized segmented images were sent to 1D Log Gabor wavelets for template generation through encoding. All generated templates were stored in template database. One template from the stored templates of each individual cattle was selected as the query template and matched with remaining templates of that individual. The findings revealed with a clear discrimination between individuals, with intra-class matching percentages consistently above 52% and self-matching reached 100% while inter-class matching values remained below 52%.The result obtained in this paper revealed that the individual cattle can be successfully identified using iris. The technology presented in this paper may be used as a tool for iris image based individual cattle identification.