Classification of Vegetables using TensorFlow

Abstract

Recognition is one of the main areas in computer vision, it yields high-level understanding by computers, one of the most important areas in recognition is object recognition which is the process of finding a specific object in an image or video sequence [16]. This paper is purposing the glimpse of the recognition of a particular vegetable [17]. This is being implemented on the TensorFlow platform, which is making use of OpenCV as the main library database. TensorFlow [20] algorithm uses tensor as its basic unit of information. Firstly the given frame is converted into an image and differentiated into cubical parts from which the features are extracted, so to converged it into the data set [3]. Such data sets are encapsulated from every cubical unit, emerged as the whole bunch of values after traversing thoroughly through given frame. Having these values, the certain frame is categorized into one of the sets of images provided, at the conclusion side percentage-wise isolation of objects is done, and here the vegetables are being identified and corresponding action should be executed [6]. Highlighting the uniqueness of usage of this idea, would result into involution of more prominent ways of segregation of vegetables in a food production industry as per the requirement, in the territory wherein the only agriculture holds the backbone of economy [18].

Deep Learning based Image Processing Solutions in Food Engineering: A Review

“…Also, a comparison of the experiment was made with the traditional BP neural network and the SVM classifier where AlexNet shows a higher classification accuracy rate of 92.1%. In (Patil and Gaikwad, 2018) deep learning technique has been used for vegetable classification. The authors perform transfer learning using Inception v3 model for extracting 205 features and finally classifying the vegetables.…”

Prototipo de un sistema que determine el estado de madurez de un plátano utilizando Deep Learning y Visión Artificial

“…Investigadores del Instituto Tecnológico de Vishwakarma en Pune desarrollaron un sistema de clasificación de vegetales a partir de imágenes, usando Tensorflow para el aprendizaje automático y OpenCV para el procesamiento de imágenes [6]. Para el modelo de red neuronal, utilizaron Inception-v3, la cual es una red neuronal profunda brindada por Tensorflow, que tomó un largo periodo de tiempo para ser entrenada.…”