scispace - formally typeset
Search or ask a question
Author

Vipul Salunke

Bio: Vipul Salunke is an academic researcher from Massachusetts Institute of Technology. The author has contributed to research in topics: Deep learning & Statistical classification. The author has an hindex of 1, co-authored 1 publications receiving 4 citations.

Papers
More filters
Book ChapterDOI
01 Jan 2021
TL;DR: It is indicative that a combination of fast recurrent neural networks and CNN may produce high accuracy with minimum time complexity, as existing researchers reflect CNN provides around 96.50% average accuracy for sentiment classification on the flicker image dataset.
Abstract: Determining the image sentiment is a tedious task for classification algorithms, owing to complexities in the raw images as well as the intangible nature of human sentiments. Classifying image sentiments is an evergreen research area, especially in social data analytics. In current times, it is a common practice for majority people to precise their feelings on the web by substituting text with the upload of images via a multiplicity of social media sites like Facebook, Instagram, Twitter as well as any other platform. To identify the emotions from visual cues, some visual features as well as image processing techniques are used. Several existing systems have already introduced emotion detection using machine learning techniques, but the traditional feature extraction strategies do not achieve the required accuracy on random objects. In the entire process, normalization of image, feature extraction, and feature selection are important tasks in the train module. This work articulates the newest developments in the field of image sentiment employing deep learning techniques. Also, the use of conventional machine learning techniques is compared along with deep learning algorithms. It is indicative that a combination of fast recurrent neural networks and CNN may produce high accuracy with minimum time complexity. It is noted from the survey that existing researchers reflect CNN provides around 96.50% average accuracy for sentiment classification on the flicker image dataset.

10 citations


Cited by
More filters
Proceedings ArticleDOI
13 May 2021
TL;DR: In this article, an automatic visual sentiment analysis (VSA) model using an optimization-based support vector machine (SVM) was developed, in which the input images' features were extracted from the weighed-FC8 layer of the pre-trained ResNet-18, where the relief algorithm evaluates the updated weight.
Abstract: This research aims to develop an automatic visual sentiment analysis (VSA) model using an optimization-based support vector machine (SVM). Initially, the input images' features are extracted from the weighed-FC8 layer of the pre-trained ResNet-18, where the relief algorithm evaluates the updated weight. On the other hand, the SVM classifier is tuned optimally using a hybrid optimization technique called Holoentropy Life Choice Optimization (HELMCO) algorithm. HELMCO has the characteristic features of both the Life Choice Based Optimization (LCBO) algorithm and the Cross entropy (CE) method. The analysis is done using the Emotion-6 and Abstract Art_photo datasets based on performance parameters, such as Accuracy, Sensitivity, and Specificity. The accuracy of the proposed model is 70.7% using the Emotion-6 dataset and 76.8% using the Art_photo Dataset.

8 citations

Book ChapterDOI
01 Jan 2021
TL;DR: This work retrieved real-time twitter data pertaining to three currently popular hashtags in the Indian context and carried out extensive experimentation analysis about the prevailing sentiment of a strata of population.
Abstract: Twitter analytics is a classic research area especially with the widespread presence of Big Data in various online media such as—social network sites, online portals for shopping, e-commerce, forums, chats, recommendation systems, and online services. Ascertaining the sentiment behind, the various types of tweets by different persons can provide great insights on various aspects including behavioral patterns. Besides highlighting the newest trends in the field, we retrieved real-time twitter data pertaining to three currently popular hashtags in the Indian context and carried out extensive experimentation analysis about the prevailing sentiment of a strata of population. Inclusion of current challenges, future trends and applications of sentiment analysis from Twitter data makes this novel work very useful for fellow researchers.

2 citations

Proceedings ArticleDOI
08 Jul 2021
TL;DR: In this article, a novel deep learning model, Long Short-Term Memory and Recurrent Neural Network (LSTM-RNN) was used to classify text emotions into multiple classes and categorize text for better accuracy.
Abstract: Emotion analysis plays a part in understanding the feelings of human beings. People's actions and speech express various feelings, behaviors and emotions which can have various impacts. Emotion and sentiment analysis is a broad research area for finding emotion which helps getting useful insight through text and speech. In most of previous work, nearly all projects have focused on analyzing the expression based on positive, negative and neutral classification. This research work analyzes the proposed system by categorizing the text into emotion classes called joy, sadness, anger, fear, love and surprise. This work helps us to label text emotions into multiple classes and categorize the text for better accuracy. This research work represents the enhancement of a novel Deep Learning model scheme, Long Short-Term Memory and Recurrent Neural Network which discuss various categories distribution on knowledgeable data. This will also summarize the previous works done on textual emotional classification based on various sentiment models and approach with comparative survey analysis.

1 citations

Proceedings ArticleDOI
08 Jul 2022
TL;DR: In this paper , an image sentiment prediction model is built using Convolutional Neural Networks(CNN) to perform sentiment classification efficiently and enhance the accuracy of restaurant image dataset posted on social media.
Abstract: In the recent years online reviews are prevalent. Over the years people have started giving feedback about a restaurant by posting images as part of a review where the sentiment polarity is classified based on the facial expressions or the foods. Even more to it is a piece of text along with the image that gives more clear understanding about the picture. As there is tremendous work carried over on text sentiment analysis(SA), in this paper we are focusing on visual analysis to identify whether a given image expresses positive or negative sentiment. In this paper, an image sentiment prediction model is built using Convolutional Neural Networks(CNN). The objective of this work is to perform sentiment classification efficiently and enhance the accuracy of restaurant image dataset posted on social media. The results show that the proposed model achieves better performance on analysis of opinions from images compared to naive bayes which is a machine learning technique.
Proceedings ArticleDOI
17 Jul 2022
TL;DR: In this paper , a neural network model is proposed to extract the features automatically and classify the vegetation's sentiment and purposes, and the obtained experimental results are evaluated with the established criteria.
Abstract: Automated time series vegetation data classification has received much attention in recent years. It has important benefits in remote sensing (RS) field such as land cover vegetation quality and productivity estimation. Recently, deep learning techniques have achieved increasing success in RS data classification including vegetation mapping. However, these approaches do not consider the vegetation's characteristics such as effective rate of crop utilization that can be useful in the process of vegetation classification. Motivated by this issue and for easy multi-temporal vegetation analysis, mapping and monitoring, in this paper; a neural network model is proposed to extract the features automatically and classify the vegetation's sentiment and purposes. We conduct experiments using multi-temporal public available Sentinel-2A datasets and the obtained experimental results are evaluated with the established criteria.