-
Analysis of Convolutional Decoder for Image Caption Generation
Authors:
Sulabh Katiyar,
Samir Kumar Borgohain
Abstract:
Recently Convolutional Neural Networks have been proposed for Sequence Modelling tasks such as Image Caption Generation. However, unlike Recurrent Neural Networks, the performance of Convolutional Neural Networks as Decoders for Image Caption Generation has not been extensively studied. In this work, we analyse various aspects of Convolutional Neural Network based Decoders such as Network complexi…
▽ More
Recently Convolutional Neural Networks have been proposed for Sequence Modelling tasks such as Image Caption Generation. However, unlike Recurrent Neural Networks, the performance of Convolutional Neural Networks as Decoders for Image Caption Generation has not been extensively studied. In this work, we analyse various aspects of Convolutional Neural Network based Decoders such as Network complexity and depth, use of Data Augmentation, Attention mechanism, length of sentences used during training, etc on performance of the model. We perform experiments using Flickr8k and Flickr30k image captioning datasets and observe that unlike Recurrent Neural Network based Decoder, Convolutional Decoder for Image Captioning does not generally benefit from increase in network depth, in the form of stacked Convolutional Layers, and also the use of Data Augmentation techniques. In addition, use of Attention mechanism also provides limited performance gains with Convolutional Decoder. Furthermore, we observe that Convolutional Decoders show performance comparable with Recurrent Decoders only when trained using sentences of smaller length which contain up to 15 words but they have limitations when trained using higher sentence lengths which suggests that Convolutional Decoders may not be able to model long-term dependencies efficiently. In addition, the Convolutional Decoder usually performs poorly on CIDEr evaluation metric as compared to Recurrent Decoder.
△ Less
Submitted 8 March, 2021;
originally announced March 2021.
-
Comparative evaluation of CNN architectures for Image Caption Generation
Authors:
Sulabh Katiyar,
Samir Kumar Borgohain
Abstract:
Aided by recent advances in Deep Learning, Image Caption Generation has seen tremendous progress over the last few years. Most methods use transfer learning to extract visual information, in the form of image features, with the help of pre-trained Convolutional Neural Network models followed by transformation of the visual information using a Caption Generator module to generate the output sentenc…
▽ More
Aided by recent advances in Deep Learning, Image Caption Generation has seen tremendous progress over the last few years. Most methods use transfer learning to extract visual information, in the form of image features, with the help of pre-trained Convolutional Neural Network models followed by transformation of the visual information using a Caption Generator module to generate the output sentences. Different methods have used different Convolutional Neural Network Architectures and, to the best of our knowledge, there is no systematic study which compares the relative efficacy of different Convolutional Neural Network architectures for extracting the visual information. In this work, we have evaluated 17 different Convolutional Neural Networks on two popular Image Caption Generation frameworks: the first based on Neural Image Caption (NIC) generation model and the second based on Soft-Attention framework. We observe that model complexity of Convolutional Neural Network, as measured by number of parameters, and the accuracy of the model on Object Recognition task does not necessarily co-relate with its efficacy on feature extraction for Image Caption Generation task.
△ Less
Submitted 23 February, 2021;
originally announced February 2021.
-
Image Captioning using Deep Stacked LSTMs, Contextual Word Embeddings and Data Augmentation
Authors:
Sulabh Katiyar,
Samir Kumar Borgohain
Abstract:
Image Captioning, or the automatic generation of descriptions for images, is one of the core problems in Computer Vision and has seen considerable progress using Deep Learning Techniques. We propose to use Inception-ResNet Convolutional Neural Network as encoder to extract features from images, Hierarchical Context based Word Embeddings for word representations and a Deep Stacked Long Short Term M…
▽ More
Image Captioning, or the automatic generation of descriptions for images, is one of the core problems in Computer Vision and has seen considerable progress using Deep Learning Techniques. We propose to use Inception-ResNet Convolutional Neural Network as encoder to extract features from images, Hierarchical Context based Word Embeddings for word representations and a Deep Stacked Long Short Term Memory network as decoder, in addition to using Image Data Augmentation to avoid over-fitting. For data Augmentation, we use Horizontal and Vertical Flip** in addition to Perspective Transformations on the images. We evaluate our proposed methods with two image captioning frameworks- Encoder-Decoder and Soft Attention. Evaluation on widely used metrics have shown that our approach leads to considerable improvement in model performance.
△ Less
Submitted 22 February, 2021;
originally announced February 2021.
-
Shed More Light on Bloom Filter's Variants
Authors:
Ripon Patgiri,
Sabuzima Nayak,
Samir Kumar Borgohain
Abstract:
Bloom Filter is a probabilistic membership data structure and it is excessively used data structure for membership query. Bloom Filter becomes the predominant data structure in approximate membership filtering. Bloom Filter extremely enhances the query response time, and the response time is very fast. Bloom filter (BF) is used to detect whether an element belongs to a given set or not. The Bloom…
▽ More
Bloom Filter is a probabilistic membership data structure and it is excessively used data structure for membership query. Bloom Filter becomes the predominant data structure in approximate membership filtering. Bloom Filter extremely enhances the query response time, and the response time is very fast. Bloom filter (BF) is used to detect whether an element belongs to a given set or not. The Bloom Filter returns True Positive (TP), False Positive (FP), or True Negative (TN). The Bloom Filter is widely adapted in numerous areas to enhance the performance of a system. In this paper, we present a) in-depth insight on the Bloom Filter,and b) the prominent variants of the Bloom Filters.
△ Less
Submitted 17 March, 2019;
originally announced March 2019.
-
scaleBF: A High Scalable Membership Filter using 3D Bloom Filter
Authors:
Ripon Patgiri,
Sabuzima Nayak,
Samir Kumar Borgohain
Abstract:
Bloom Filter is extensively deployed data structure in various applications and research domain since its inception. Bloom Filter is able to reduce the space consumption in an order of magnitude. Thus, Bloom Filter is used to keep information of a very large scale data. There are numerous variants of Bloom Filters available, however, scalability is a serious dilemma of Bloom Filter for years. To s…
▽ More
Bloom Filter is extensively deployed data structure in various applications and research domain since its inception. Bloom Filter is able to reduce the space consumption in an order of magnitude. Thus, Bloom Filter is used to keep information of a very large scale data. There are numerous variants of Bloom Filters available, however, scalability is a serious dilemma of Bloom Filter for years. To solve this dilemma, there are also diverse variants of Bloom Filter. However, the time complexity and space complexity become the key issue again. In this paper, we present a novel Bloom Filter to address the scalability issue without compromising the performance, called scaleBF. scaleBF deploys many 3D Bloom Filter to filter the set of items. In this paper, we theoretically compare the contemporary Bloom Filter for scalability and scaleBF outperforms in terms of time complexity.
△ Less
Submitted 15 March, 2019;
originally announced March 2019.
-
Role of Bloom Filter in Big Data Research: A Survey
Authors:
Ripon Patgiri,
Sabuzima Nayak,
Samir Kumar Borgohain
Abstract:
Big Data is the most popular emerging trends that becomes a blessing for human kinds and it is the necessity of day-to-day life. For example, Facebook. Every person involves with producing data either directly or indirectly. Thus, Big Data is a high volume of data with exponential growth rate that consists of a variety of data. Big Data touches all fields, including Government sector, IT industry,…
▽ More
Big Data is the most popular emerging trends that becomes a blessing for human kinds and it is the necessity of day-to-day life. For example, Facebook. Every person involves with producing data either directly or indirectly. Thus, Big Data is a high volume of data with exponential growth rate that consists of a variety of data. Big Data touches all fields, including Government sector, IT industry, Business, Economy, Engineering, Bioinformatics, and other basic sciences. Thus, Big Data forms a data silo. Most of the data are duplicates and unstructured. To deal with such kind of data silo, Bloom Filter is a precious resource to filter out the duplicate data. Also, Bloom Filter is inevitable in a Big Data storage system to optimize the memory consumption. Undoubtedly, Bloom Filter uses a tiny amount of memory space to filter a very large data size and it stores information of a large set of data. However, functionality of the Bloom Filter is limited to membership filter, but it can be adapted in various applications. Besides, the Bloom Filter is deployed in diverse field, and also used in the interdisciplinary research area. Bioinformatics, for instance. In this article, we expose the usefulness of Bloom Filter in Big Data research.
△ Less
Submitted 15 March, 2019;
originally announced March 2019.
-
Preventing DDoS using Bloom Filter: A Survey
Authors:
Ripon Patgiri,
Sabuzima Nayak,
Samir Kumar Borgohain
Abstract:
Distributed Denial-of-Service (DDoS) is a menace for service provider and prominent issue in network security. Defeating or defending the DDoS is a prime challenge. DDoS make a service unavailable for a certain time. This phenomenon harms the service providers, and hence, loss of business revenue. Therefore, DDoS is a grand challenge to defeat. There are numerous mechanism to defend DDoS, however,…
▽ More
Distributed Denial-of-Service (DDoS) is a menace for service provider and prominent issue in network security. Defeating or defending the DDoS is a prime challenge. DDoS make a service unavailable for a certain time. This phenomenon harms the service providers, and hence, loss of business revenue. Therefore, DDoS is a grand challenge to defeat. There are numerous mechanism to defend DDoS, however, this paper surveys the deployment of Bloom Filter in defending a DDoS attack. The Bloom Filter is a probabilistic data structure for membership query that returns either true or false. Bloom Filter uses tiny memory to store information of large data. Therefore, packet information is stored in Bloom Filter to defend and defeat DDoS. This paper presents a survey on DDoS defending technique using Bloom Filter.
△ Less
Submitted 15 October, 2018;
originally announced October 2018.
