Recent deep learning methods for fruits classification resulted in promising performance. However, these methods are with heavy-weight architectures in nature, and hence require a higher storage and expensive training operations due to feeding a large number of training parameters. There is a necessity to explore lightweight deep learning models without compromising the classification accuracy. In this paper, we propose a lightweight deep learning model using the pre-trained MobileNetV2 model and attention module. First, the convolution features are extracted to capture the high-level object-based information. Second, an attention module is used to capture the interesting semantic information. The convolution and attention modules are then combined together to fuse both the high-level object-based information and the interesting semantic information, which is followed by the fully connected layers and the softmax layer. Evaluation of our proposed method, which leverages transfer learning approach, on three public fruit-related benchmark datasets shows that our proposed method outperforms the four latest deep learning methods with a smaller number of trainable parameters and a superior classification accuracy. Our model has a great potential to be adopted by industries closely related to the fruit growing and retailing or processing chain for automatic fruit identification and classifications in the future.

/pdf/fruit-classification-using-attention-based-mobilenetv2-for-312thd9x.pdf

Fruit classification using attention-based MobileNetV2 for industrial applications

/pdf/gan-computers-generate-arts-a-survey-on-visual-arts-music-1i2utbgn.pdf

GAN computers generate arts? A survey on visual arts, music, and literary text generation using generative adversarial network

"Art is the lie that enables us to realize the truth." - Pablo Picasso. For centuries, humans have dedicated themselves to producing arts to convey their imagination. The advancement in technology and deep learning in particular, has caught the attention of many researchers trying to investigate whether art generation is possible by computers and algorithms. Using generative adversarial networks (GANs), applications such as synthesizing photorealistic human faces and creating captions automatically from images were realized. This survey takes a comprehensive look at the recent works using GANs for generating visual arts, music, and literary text. A performance comparison and description of the various GAN architecture are also presented. Finally, some of the key challenges in art generation using GANs are highlighted along with recommendations for future work.

GAN Computers Generate Arts? A Survey on Visual Arts, Music, and Literary Text Generation using Generative Adversarial Network.

Longan fruit (Dimocarpus longan) originating from Southeast Asia, is a fruit that is commonly consumed in the region. The fruit is susceptible to damage during the distribution process. The selection of fruit that is suitable for consumption and that which is not is still made manually. This study aims to build a model using machine learning that can automatically classify longan edibility. Two classification algorithms are used, namely the k-nearest neighbor and the support vector machine. The model is trained with a number of images consisting of two classes, namely, fit (accepted) and unfit (rejected) to eat. The experimental results show that the algorithm's accuracy for k- nearest neighbors and the support vector machine are 93% and 98%, respectively. These results indicate that the classification of longan using machine learning is quite promising.

Classification of Longan Edibility using Machine Learning

Cross-domain image retrieval: methods and applications

Fruit classification plays an important role in many industrial applications including factories, supermarkets and other fields. The importance of fruit classification can also be seen among people with dietary requirements to assist them in selecting the correct categories of fruits. Manual sorting method for fruit classification was used earlier which is time consuming and involves human presence always. Many machine learning approaches for fruit classification were proposed in the past. Deep learning can be a powerful engine for producing actionable results in today’s world with its abilities of recognition and classification. Thus, an efficient model for fruit classification was developed using convolutional neural network in deep learning. It uses fruits 360 dataset containing 131 different classes of fruits and vegetables that can classify images into different categories. The model was developed using TensorFlow backend. It was trained using 50 epochs, and it obtained an accuracy of 94.35%.

Fruit Classification Using Deep Learning

Sketch based image retrieval (SBIR) is a methodology in which the system can recognize an image and retrieve from the database, given an input query sketch. Socio economic importance and relevance of Sketch based image retrieval systems are high as the users of touch screen based devices are rapidly increasing even in remote areas. The main challenge of this task is that sketches contain only an abstract visual information compared to images. The proposed method uses Siamese CNN for the implementation of SBIR. Siamese CNN joins two convolutional neural networks with one loss function. The transfer learning approach is implemented to learn the feature space of both sketch input query and the images. The VGG19 pre-trained model is used to obtain feature descriptors of sketches and images. The dataset Sketchy Database containing 75,471 sketches of 125 categories is used. The results obtained shows that proposed method gives better performance compared to the similar methods.

Cross Domain Descriptor for Sketch Based Image Retrieval using Siamese network

Today, Technology have remarkable charm in the area of Computer Graphics and Vision. Producing absolute images from the poor hand-drawn sketches is a very demanding and laborious task in this area. Hand-drawn sketch recognition is widely used in sketch based image and video retrieval, manipulations and reorganizations. In S ketch to image synthesis, the sketches are translated to realistic images with the use of a generative model. An image is put forward to image translation network that involves in producing a synthesized image from the input sketch via an adversarial process. A novel Conditional Generative Adversarial Network (cGANs) which is an extension of Generative Adversarial Networks (GANs) is used to produce the images with some sort of conditions or attributes. In this work, the implementation of cGANs for synthesizing the images from hand-drawn sketches gives a remarkable output. The performance of the proposed sketch to image translation network was excellent and appreciable.

Synthesizing Images from Hand-Drawn Sketches using Conditional Generative Adversarial Networks

Sketch based image recognition is an approach where the system recognizes an image from the database based on the query sketch received from the user through an interface. In our implementation, we use transfer learning approach to learn the features of the sketch. Transfer learning is a technique used to reuse a model, which is pretrained, for a new problem thereby enhancing the model accuracy and reducing training time. Sketchy Database is used for training which contains 75,471 sketches of 125 categories. A pretrained VGG19 network is used to find the similarity between the sketch and image database using a cosine similarity function. Based on the result of the similarity function, SBIR recognizes a set of similar images corresponding to the sketch. Since the uses of touch devices are increasing rapidly even in a relatively small area, we focus on our final solution being scalable as well as distributed.

Veena A. Kumar

Papers

Fruit Classification Using Deep Learning

Cross Domain Descriptor for Sketch Based Image Retrieval using Siamese network

Synthesizing Images from Hand-Drawn Sketches using Conditional Generative Adversarial Networks

Sketch Based Image Retrieval using Transfer Learning