Devendra Kumar Sahu | arXiv Science

arXiv2026

SN 2023dbc in M108: Optical and Near-Infrared Observations of a Highly-Obscured, Moderately Energetic Stripped-Envelope Supernova

We present near-infrared (NIR) and optical observations of the highly reddened and moderately energetic Type Ib supernova (SN) 2023dbc, {\bf covering a period from} 2 to 136 days after the explosion. By comparing its color {\bf evolution}, specifically in $r-JHK_{\mathrm{s}}$ and $i-JHK_{\mathrm{s}}$, with those of broad-lined Type Ic (Ic-BL) and Type IIb SNe, we estimate a significant extinction of $A_{V}=4.1\pm0.1$\,mag toward the SN. The extinction-corrected peak absolute magnitudes are $M_{J} = -16.8\pm0.2$\,mag, $M_{H} = -16.8\pm0.2$\,mag, and $M_{K_{\mathrm{s}}} = -17.0\pm0.2$\,mag. The SN {\bf exhibited} an $r$-band rise time of 14.9 days. The spectra {\bf display} broad features {\bf indicative of} high expansion velocities; the He~{\sc i} line velocity was measured at $16,000\,\mathrm{km\,s^{-1}}$ at $t=-4$\,d. Its spectral profile is broader than {\bf those} of typical moderately energetic Type Ib SNe, {\bf yet narrower than those of Type Ic-BL SNe, placing it in an intermediate category}. Based on the light-curve timescale and velocity, we estimate {\bf a} kinetic energy of $E_k = (4.1\pm0.7) \times 10^{51}$\,erg, {\bf an} ejecta mass of $M_{\mathrm{ej}} = 2.3\pm0.7\,M_{\odot}$, and a radioactive $^{56}\mathrm{Ni}$ mass of $(3.8\pm0.1) \times 10^{-2}\,M_{\odot}$. {\bf An} analysis using a two-component model suggests a steep density profile in the outer layer {\bf contrasted with} a dense inner core, {\bf which implies} ejecta asphericity. The low $^{56}\mathrm{Ni}$ mass is consistent with a partial fallback scenario. We conclude that SN 2023dbc originated from an aspherical explosion with partial core fallback, {\bf arising} from a progenitor ($M_{\mathrm{ini}} \simeq 15\,M_{\odot}$) that had retained {\bf its} helium envelope {\bf within} a binary system.

astro-ph.HE↗

arXiv2015

Sequence to Sequence Learning for Optical Character Recognition

We propose an end-to-end recurrent encoder-decoder based sequence learning approach for printed text Optical Character Recognition (OCR). In contrast to present day existing state-of-art OCR solution which uses connectionist temporal classification (CTC) output layer, our approach makes minimalistic assumptions on the structure and length of the sequence. We use a two step encoder-decoder approach -- (a) A recurrent encoder reads a variable length printed text word image and encodes it to a fixed dimensional embedding. (b) This fixed dimensional embedding is subsequently comprehended by decoder structure which converts it into a variable length text output. Our architecture gives competitive performance relative to connectionist temporal classification (CTC) output layer while being executed in more natural settings. The learnt deep word image embedding from encoder can be used for printed text based retrieval systems. The expressive fixed dimensional embedding for any variable length input expedites the task of retrieval and makes it more efficient which is not possible with other recurrent neural network architectures. We empirically investigate the expressiveness and the learnability of long short term memory (LSTMs) in the sequence to sequence learning regime by training our network for prediction tasks in segmentation free printed text OCR. The utility of the proposed architecture for printed text is demonstrated by quantitative and qualitative evaluation of two tasks -- word prediction and retrieval.

cs.CV↗