Conference Papers

Permanent URI for this collectionhttps://idr.nitk.ac.in/handle/123456789/28506

Browse

Author: search.filters.author.Jeppu, N.Has files: No

Search Results

Now showing 1 - 2 of 2
  • No Thumbnail Available
    Item
    Teaching formal methods at undergraduate/graduate level: The three perspectives
    (Institute of Electrical and Electronics Engineers Inc., 2018) Jeppu, N.; Jeppu, Y.; Kavitha Devi, M.K.K.
    Formai methods provide easy way of validating properties about systems. These methods are in existence since the last 50 years but have not been used fully by the industry as an engineering tool. One of the challenges of acceptance is education. Educating engineering students to take up formal methods is a challenge. This paper looks at these aspects of formal methods by providing demonstration of its usefulness on a recent failure. A few challenges of teaching formal method are described and three perspectives of formal methods are explained. The viewpoints are from a student who has worked on this, a teacher who teaches this and an industry practitioner of formal methods. We advocate an industry academia partnership to overcome some of these challenges of teaching formal methods to students. © 2017 IEEE.
  • No Thumbnail Available
    Item
    Extending Denoising AutoEncoders for Feature Recognition
    (Institute of Electrical and Electronics Engineers Inc., 2018) Jeppu, N.; Chandrasekaran, K.
    Image rendering techniques involve the addition of noise on the rendered samples. The characteristics of Monte Carlo rendered noisy images makes it difficult to denoise using conventional methods. The noise induced in this case is spatially sparse. Using traditional methods of denoising and feature recognition is time consuming for such images as they use the entire image space as their search space. This results in unnecessary calculations that can be avoided and therefore reduce the processing time significantly. A recurrent convolutional neural network that operates on varying spatial resolutions, also known as the auto encoder decoder structure perform very well on these rendered images. The partitioning into encoder and decoder phases lets the network operate on continuously decreasing and increasing spatial domains respectively. This can be coupled with classification techniques to incorporate feature recognition of noisy images. In this paper the pretrained autoencoder layers are supported with additional softmax and sigmoid layers to enable feature recognition capabilities. © 2018 IEEE.