# ExTTNet: A Deep Learning Algorithm for Extracting Table Texts from Invoice Images

arXiv: https://arxiv.org/abs/2402.02246 · HTML: https://mlsanity.com/paper/2402.02246/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 5 of its 43 references that are in the map.
- First posted on arXiv 3 Feb 2024.

## Builds on (most cited first, 5 of 5)

- [CloudScan - A configuration-free invoice analysis system using recurrent neural networks](https://mlsanity.com/paper/1708.07403/) (2017)
- [Robust Table Detection and Structure Recognition from Heterogeneous Document Images](https://mlsanity.com/paper/2203.09056/) (2022)
- [Hyper-Realist Rendering: A Theoretical Framework](https://mlsanity.com/paper/2401.12853/) (2024)
- [GenPluSSS: A Genetic Algorithm Based Plugin for Measured Subsurface Scattering Representation](https://mlsanity.com/paper/2401.15245/) (2024)
- [MelNet: A Real-Time Deep Learning Algorithm for Object Detection](https://mlsanity.com/paper/2401.17972/) (2024)

## Built on by (most cited first, 3 of 3)

- [LightFFDNets: Lightweight Convolutional Neural Networks for Rapid Facial Forgery Detection](https://mlsanity.com/paper/2411.11826/) (2024)
- [OkanNet: A Lightweight Deep Learning Architecture for Classification of Brain Tumor from MRI Images](https://mlsanity.com/paper/2604.01264/) (2026)
- [TRACK: Telemetry-Based Racing Analysis and Coaching Kit in Sim Racing Games](https://mlsanity.com/paper/2610.10061/) (2026)

## Abstract (arXiv)

In this work, product tables in invoices are obtained autonomously via a deep learning model, which is named as ExTTNet. Firstly, text is obtained from invoice images using Optical Character Recognition (OCR) techniques. Tesseract OCR engine [37] is used for this process. Afterwards, the number of existing features is increased by using feature extraction methods to increase the accuracy. Labeling process is done according to whether each text obtained as a result of OCR is a table element or not. In this study, a multilayer artificial neural network model is used. The training has been carried out with an Nvidia RTX 3090 graphics card and taken 162 minutes. As a result of the training, the F1 score is 0.92.