Infoscribe.ai Blog

History of Computer Vision (2/3): From 1980 to the Late 1990s

computer-vision-history-1980-late-1990s

By: Mathieu G.
2023-07-19

History of Computer Vision (2/3): From 1980 to the Late 1990s

The field of computer vision has witnessed remarkable advancements since its inception. In this article, we will delve into the period from 1980 to the late 1990s, highlighting significant milestones, key players, and technological breakthroughs that shaped the evolution of computer vision during this era.

Rise of Image Processing Techniques:

In the early 1980s, image processing techniques became a prominent focus of computer vision research. Researchers, such as John Canny and David Lowe, developed algorithms for edge detection, image enhancement, and feature extraction, enabling more robust analysis of visual data.

Development of Object Recognition Algorithms:

During the 1980s, efforts were made to develop algorithms capable of recognizing and classifying objects within images. David Marr, Tomaso Poggio, and their colleagues laid the foundation for object recognition by introducing the concept of multi-scale image analysis and proposing models for object representation and inference.

Integration of Artificial Intelligence and Computer Vision:

In the late 1980s, there was a significant shift towards integrating artificial intelligence (AI) techniques with computer vision. Prominent researchers, including Geoffrey Hinton, Yann LeCun, and Terry Winograd, explored the application of neural networks and machine learning algorithms to improve object recognition, scene understanding, and visual perception.

Advancements in 3D Computer Vision:

The 1990s witnessed remarkable progress in 3D computer vision. Marc Pollefeys, Richard Hartley, and Jean Ponce made significant contributions to the field by developing techniques for 3D reconstruction from multiple images, camera calibration, and structure from motion, paving the way for applications in robotics, virtual reality, and augmented reality.

Face Recognition and Biometrics:

In the 1990s, researchers such as Takeo Kanade, Alex Pentland, and Pawan Sinha made significant breakthroughs in face recognition and biometrics. Their work laid the foundation for techniques like eigenfaces, active appearance models, and local binary patterns, which are still widely used in face recognition systems today.

Industrial Applications and Commercialization:

By the late 1990s, computer vision started making its way into various industrial applications. Companies like Cognex Corporation and Matrox Imaging introduced automated quality control systems for industries like manufacturing and electronics. Medical imaging technologies, such as MRI and CT scanners, incorporated advanced computer vision algorithms for more accurate diagnostics.

Hours

Monday - Saturday: 08:00 AM - 9:00 PM

Location

15 Av. Guy Môquet, 94340 Joinville-le-Pont,
France

Production units

Ankerana
Antananarivo,Madagascar

×
Delivering accurate and consistent image annotation services

At Infoscribe, we understand that annotating images requires a great deal of time and utmost precision. Our primary goal is to ensure the highest level of quality to achieve 100% customer satisfaction.

To achieve this goal, we have implemented a rigorous training program for our annotators. Before being assigned to a project, each annotator is trained on best practices and is given test data to ensure a thorough understanding of the project and all possible scenarios. This allows us to deliver accurate and consistent results to our clients.

×
Ensuring strict quality control: our process at infoscribe

At Infoscribe, we prioritize quality control to ensure accurate and consistent results for our clients. Here's how we do it:

1
Quality control
Before launching a project, we conduct a 100% quality control to analyze and address any frequent errors caused by misinterpretations or misunderstandings of instructions.
2
QC Reports
Our QC team creates a report for each quality control performed, listing and illustrating non-conformities detected with screenshots.
3
Corrections
Project managers use these QC reports to explain errors to annotators so they can make corrections.
4
Improvement
We also use whiteboards to communicate common errors and encourage continuous improvement of our quality.
5
Sampling inspection
Once the compliance rate is high and stable after a few weeks, we perform sampling inspection based on the ISO2859 standard (2000 version).
×
Project management



Our project managers, who are in direct contact with our customers, comply with a detailed checklist designed to prevent mistakes and they report on a daily or weekly basis depending on the needs our customers expressed at the beginning of a project.