publications

2024

1. AAAI-24

   Evaluating Pre-Trial Programs Using Interpretable Machine Learning Matching Algorithms for Causal Inference

   Saksham Jain*, Travis Seale-Carlisle* , Courtney Lee , and 6 more authors

   AAAI-24, 2024

   Abs HTML PDF Poster Slides

   

2023

1. NeurIPS-22 WS

   Closing the domain gap: blended synthetic imagery for climate object detection

   Caleb Kornfein , Saksham Jain*, Frank Willard* , and 5 more authors

   Environmental Data Science, 2023

   Abs HTML PDF Slides

   

   Accurate geospatial information about the causes and consequences of climate change, including energy systems infrastructure, is critical to planning climate change mitigation and adaptation strategies. When up-to-date spatial data on infrastructure is lacking, one approach to fill this gap is to learn from overhead imagery using deep-learning-based object detection algorithms. However, the performance of these algorithms can suffer when applied to diverse geographies, which is a common case. We propose a technique to generate realistic synthetic overhead images of an object (e.g., a generator) to enhance the ability of these techniques to transfer across diverse geographic domains. Our technique blends example objects into unlabeled images from the target domain using generative adversarial networks. This requires minimal labeled examples of the target object and is computationally efficient such that it can be used to generate a large corpus of synthetic imagery. We show that including these synthetic images in the training of an object detection model improves its ability to generalize to new domains (measured in terms of average precision) when compared to a baseline model and other relevant domain adaptation techniques.

2022

1. 

   Synthetic data augmentation for surface defect detection and classification using deep learning

   Saksham Jain, Gautam Seth , Arpit Paruthi , and 2 more authors

   Journal of Intelligent Manufacturing, 2022

   Abs HTML PDF

   

   Deep learning techniques, especially Convolutional Neural Networks (CNN), dominate the benchmarks for most computer vision tasks. These state-of-the-art results are typically obtained through supervised learning, for which large annotated datasets are required. However, acquiring such datasets for manufacturing applications remains a challenging proposition due to the time and costs involved in their collection. To overcome this disadvantage, a novel framework is proposed for data augmentation by creating synthetic images using Generative Adversarial Networks (GANs). The generator synthesizes new surface defect images from random noise which is trained over time to get realistic fakes. These synthetic images can be used further for training of classification algorithms. Three GAN architectures are trained, and the entire data augmentation pipeline is implemented for the Northeastern University (China) Classification (NEU-CLS) dataset for hot-rolled steel strips from NEU Surface Defect Database. The classification accuracy of a simple CNN architecture is measured on synthetic augmented data and further it is compared with similar state-of-the-arts. It is observed that the proposed GANs-based augmentation scheme significantly improves the performance of CNN for classification of surface defects. The classically augmented CNN yields sensitivity and specificity of 90.28% and 98.06% respectively. In contrast, the synthetically augmented CNN yields better results, with sensitivity and specificity of 95.33% and 99.16% respectively. Also, the use of GANs is demonstrated to disentangle the representation space and to add additional domain knowledge through synthetic augmentation that can be difficult to replicate through classic augmentation. The proposed framework demonstrates high generalization capability. It may be applied to other supervised surface inspection tasks, and thus facilitate the development of advanced vision-based inspection instruments for manufacturing applications.

2021

1. IEEE TPAMI

   Mining data impressions from deep models as substitute for the unavailable training data

   Gaurav Kumar Nayak , Konda Reddy Mopuri , Saksham Jain, and 1 more author

   IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021

   Abs HTML PDF

   

   Pretrained deep models hold their learnt knowledge in the form of model parameters. These parameters act as “memory” for the trained models and help them generalize well on unseen data. However, in absence of training data, the utility of a trained model is merely limited to either inference or better initialization towards a target task. In this paper, we go further and extract synthetic data by leveraging the learnt model parameters. We dub them Data Impressions , which act as proxy to the training data and can be used to realize a variety of tasks. These are useful in scenarios where only the pretrained models are available and the training data is not shared (e.g., due to privacy or sensitivity concerns). We show the applicability of data impressions in solving several computer vision tasks such as unsupervised domain adaptation, continual learning as well as knowledge distillation. We also study the adversarial robustness of lightweight models trained via knowledge distillation using these data impressions. Further, we demonstrate the efficacy of data impressions in generating data-free Universal Adversarial Perturbations (UAPs) with better fooling rates. Extensive experiments performed on benchmark datasets demonstrate competitive performance achieved using data impressions in absence of original training data.

2020

1. 

   Fusion of Deep and Non-Deep Methods for Fast Super-Resolution of Satellite Images

   Saksham Jain*, Gaurav Kumar Nayak* , R Venkatesh Babu , and 1 more author

   In 2020 IEEE Sixth International Conference on Multimedia Big Data (BigMM) , 2020

   Abs HTML PDF

   

   In the emerging commercial space industry there is a drastic increase in access to low cost satellite imagery. The price for satellite images depends on the sensor quality and revisit rate. This work proposes to bridge the gap between image quality and the price by improving the image quality via super-resolution (SR). Recently, a number of deep SR techniques have been proposed to enhance satellite images. However, none of these methods utilize the region-level context information, giving equal importance to each region in the image. This, along with the fact that most state-of-the-art SR methods are complex and cumbersome deep models, the time taken to process very large satellite images can be impractically high. We, propose to handle this challenge by designing an SR framework that analyzes the regional information content on each patch of the low-resolution image and judiciously chooses to use more computationally complex deep models to super-resolve more structure-rich regions on the image, while using less resource-intensive non-deep methods on non-salient regions. Through extensive experiments on a large satellite image, we show substantial decrease in inference time while achieving similar performance to that of existing deep SR methods over several evaluation measures like PSNR, MSE and SSIM.

2019

1. 

   Robust Detection of Iris Region Using an Adapted SSD Framework

   Saksham Jain, and Indu Sreedevi

   In Computer Vision Applications: Third Workshop, WCVA 2018, Held in Conjunction with ICVGIP 2018, Hyderabad, India, December 18, 2018, Revised Selected Papers 3 , 2019

   Abs HTML PDF

   

   Accurate detection of the iris is a crucial step in several biometric tasks, such as iris recognition and spoofing detection, among others. In this paper, we consider the detection task to be the delineation of the smallest square bounding box that surrounds the iris region. To overcome the various challenges of the iris detection task, we present an efficient iris detection method that leverages the SSD (Single Shot multibox Detector) model. The architecture of SSD is modified to give a lighter and simpler framework capable of performing fast and accurate detection on the relatively smaller sized iris biometric datasets. Our method is evaluated on 4 datasets taken from different biometric applications and from the literature. It is also compared with baseline methods, such as Daugman’s algorithm, HOG+SVM and YOLO. Experimental results show that our modified SSD outperforms these other techniques in terms of speed and accuracy. Moreover, we introduce our own near-infrared image dataset for iris biometric applications, containing a robust range of samples in terms of age, gender, contact lens presence, and lighting conditions. The models are tested on this dataset, and shown to generalise well. We also release this dataset for use by the scientific community.

2018

1. 

   A cascade predictive control strategy for active suspension systems

   Shrey Jain , Saksham Jain, and Abhay Prithvi Komanduri

   In 2018 IEEE International Conference on Automation/XXIII Congress of the Chilean Association of Automatic Control (ICA-ACCA) , 2018

   Abs HTML PDF

   

   In this paper, a non-linear Active Suspension System (ASS) for a quarter car model is being controlled using a Neural Network Predictive Controller (NNPC). The non-linearities are being introduced in the system through the actuator, which is being controlled using a simple PI controller. The design of the system consists of two loops. The inner loop is the force control of the actuator, while the outer loop controls the displacement of the system when subjected to road disturbances. The gains of PI controller are optimized using Simulated Annealing (SA). Simulations are carried out in Matlab/Simulink, with desired output showing effectiveness of the proposed control system.