基于SSD网络的宠物狗检测与分类

Detection and classification of pet dogs based on SSD

目标检测作为深度学习中的重要技术已经被广泛的应用于各个方面。基于深度学习的目标检测相关技术也逐渐成熟。由于人们生活水平的提高,猫、狗等宠物受到大多数人的青睐。宠物种类的复杂性是如今面临的难题。为了提高辨别宠物种类的能力,本文以常见宠物种类--狗做为研究对象,提出了一种基于SSD(Single Shot MultiBox Detector)模型的宠物狗种类检测与分类,通过建立宠物狗的数据集,然后进行模型训练。实验证明,该方法能有效地实现对宠物狗的检测和分类。

As an important technology in deep learning, object detection has been widely applied in various aspects. Object detection technology based on deep learning is also gradually mature. With the improvement of people’s living standard, cats, dogs and other pets are favored by most people. The complexity of pet species is a challenge today. In order to improve the ability to identify pet species, this paper took the common pet species--dog as the research object and proposed a kind of pet dog species detection and classification based on SSD(Single Shot MultiBox Detector) model. Through the establishment of the data set of pet dogs, model training was carried out. Experiments show that this method can effectively detect and classify pet dogs.