New Infographic: Social Media Monitoring, don’t miss the next step!
If you want know more about Visual Brand Intelligence, download our whitepaper “Brand Intelligence in the age of Imagery”.
The Biggest Hole In Your Content Marketing Strategy: Images
by Stephen Shepherd, General Manager of LTU Technologies Inc.
Gone are the days when it was enough for a brand to just have a Facebook page and a Twitter feed. Now, brands have amassed thousands — if not millions — of followers on these and other social platforms, and these followers won’t come back unless brands provide great content. Social media marketing has morphed into “content marketing.”
AgileZen Beta User Interview: LTU Technologies
On a interview given to AgileZen, Bertrand Vidal, Software Engineer at LTU Technologies, explains How does LTU use AgileZen, and why?
To read the full story click here.
5 tips for creating a branded social image strategy.
When Facebook recently announced that its newest update would put a greater emphasWhen Facebook recently announced that its newest update would put a greater emphasis on images, the migration from a text-based to a visual social web gathered steam. Already, Pinterest, Instagram, Tumblr, and other image-driven social sites have attracted millions of users engaged in posting, pinning, and browsing colorful and catchy images.
By Stephen Shepherd - General Manager, LTU Technologies Inc.
Article published on iMedia Connection.
Read more.
How to build a social image strategy ?
As brands lose control over the dissemination of one of their most valuable strategic assets, we may think it’s a negative phenomenon, but in reality this is a huge opportunity to generate new revenues, increase loyalty and strengthen brand value and operations. Kristina: The social space is exploding. How can a brand take all of the random instances of consumers using their images from random to a strategy?
by Kristina Knight
Read more on BizReport
Big Data Helps Brands Take On Logo Photoshoppers
InformationWeek.com. LTU Technologies merges image recognition technology with analytics to help companies find and stop improper logo use.
By Jeff Bertolucci.
Read more on InformationWeek.
How to control images in the social space?
As the social space grows the issues surrounding proper branding of a product or, well, brand continue to evolve. While most uses of a brand logo are innocuous, there are times when a branded image could be placed near content that could be harmful.
by Kristina Knight
Read more on BizReport.
FINDUS fait “aussi” le Buzz Visuel sur Twitter !
Deux visuels tiennent le haut du pavé !
Une course folle de chevaux suivie de près par la plus célèbre des belges, la bien nommée Martine.
Parmi les autres détournements identifiés dans les 12 dernières heures:
Cette analyse visuelle a été effectuée sur les flux visuels Twitter des dernières 12H grâce aux algorithmes d’images de LTU technologies.
Pour en savoir plus sur la surveillance de marque sur les réseaux sociaux, téléchargez notre livre blanc. Cliquez-ici.
Deep Learning programs create the Buzz!
With an article published last week by the New York Times, artificial neural networks, also called Deep Learning programs, hit again the scientific headlines, especially in the Image Recognition area.
What are they?
Artificial neural networks, an idea going back to the 1950s, seek to mimic theway the brain absorbs information and learns from it. From that date, the technique has been applied in fields as diverse as computer vision (Image Recognition), speech recognition and the identification of promising new molecules for designing drugs, with more or less success especially in the image recognition area.
What’s new?
With the improvement of computing power and the use of GPU’s to train a model from a large set of data, Deep Learning approaches now register stunning results in terms of speed and accuracy, opening new opportunities for scientists and companies. The recent results are so impressive that some scientists even claim that this technique is the future of Silicon Valley technologies.
The technique won several scientific contests in recent months.
In October, a team of graduate students studying with the University of Toronto computer scientist Geoffrey E. Hinton won the top prize in a contest (kaggle’s challenge) sponsored by Merck to design software to help find molecules that might lead to new drugs. From a data set describing the chemical structure of 15 different molecules, they used deep-learning software to determine which molecule was most likely to be an effective drug agent. See the details results of the contest here, and an interview of the winner there.
At the same time, the same technique won an ECCV contest – the event we talked about few weeks ago - aiming to classify a large set of visuals. The technique impressed for two main reasons:
- It registered 10 points of errors fewer than the second method based on actual computer vision used today,
- But mainly, it succeeded in learning the model from a set of 1.5 million visuals in less than 3 days, opening new opportunities in the visual classification field. See the presentation here.
What does that mean for you?
In the future, as mentioned by Venture Beat, science writer John Markoff posits that deep learning will make surveillance technologies cheaper and more accessible, help marketers comb through data to identify consumer buying patterns, save even more time & money to automatically classify large set of visuals, and may also pave the way for self-driving cars and robots that can replace human workers.
To learn more about Deep Learning programs:
- Wikipedia
- A post from David Lowe – SIFT inventor
- New York Times article : Scientists See Promise in Deep-Learning Programs
- A talk dedicated to Deep Learning
- Contact Us.
Blippar unveils positive results on the effectiveness of Interactive magazines.
Why will you make your magazine interactive? If you still hesitate, the following computer graphic published by Blippar end of last week, should make you change your mind.
To see Blippar in action, download first its mobile App, then, click here.
About Blippar: Blippar™ is the first image-recognition phone app aimed at bringing to life real-world newspapers, magazines, products and posters with exciting augmented reality experiences and instantaneous content. The company launched in the UK in the Summer of 2011 and will be expanding globally throughout 2012.
Quels services digitaux développer sur les points de vente ?
«Avec la montée en puissance du e-commerce, et plus généralement du digital, le commerce en point de vente va plus évoluer dans les dix prochaines années que sur l’ensemble du siècle passé » déclare Xavier Baudoin, senior manager digital e-commerce chez BearingPoint.
Parmi les points d’amélioration listés dans leur étude, BearingPoint et le CNCC (Conseil National des centres Commerciaux) suggèrent de penser “Hub digital” ou encore de prolonger la relation via la mobile.
Pour en savoir plus sur cette étude, consultez l’article publié par LSA, sous le titre: “Les centres commerciaux envisagent leur mue sous l’effet du e-commerce”. Vous y découvrirez de nouvelles pistes de réflexion et les technologies qui vont révolutionner dès demain nos points de vente.
Et pour découvrir concrètement en quoi la reconnaissance d’image participe de cette révolution, contactez-nous au +33 1 53 43 01 68 ou demandez une démo via notre formulaire de contact.
Image Recognition : a “third wave” shopping technology for e-Bay
According to e-Bay, third wave shopping technologies - Image Recognition, Interactive TV, In-Store technologies, Augmented reality and Smart Devices - will drive £2.4bn sales growth for the UK retail sector by 2014.
85 percent of consumers buy products based on color
Color is an extremely important part of marketing and advertising. In an article published few weeks ago, the journalist Regina Woods stated that 85% of shoppers buy a particular product because of its color.
Find out more statistics and meanings behind different colors in this graphic:
Full article available on Ragan.com
Return From ECCV : one of the top 3 computer vision conferences worldwide.
As we mentioned earlier, The European Conference on Computer Vision (ECCV 2012) was held on 7-13 October, and we were there! Over the years, ECCV has become one of the top 3 computer vision conferences worldwide. The number of submitted papers has grown by 22% since 2010, up to 1437 this year. The fact that the conference was held in beautiful Florence didn’t hurt either :)
The conference crew deserve a special mention, as they pulled off a flawlessly organized conference and outstanding evening special events. The presence of a Twitter board in the main conference hall showed that the academic world is (finally) embracing the social movement — and that’s something that we like, too! Among the research works we enjoyed most, our personal “Best Paper Award” would go to Semantic Segmentation with Second-Order Pooling, by Carreira et al. What we loved in this work was both the simplicity of the method and the quality of the segmentation results. It seems that semantic segmentation has now reached a production-level of quality and speed. Other great papers which received conference awards are listed here.
But the presentation that absolutely blew our minds was not part of the main conference: it was the presentation of the results from one of the teams competing in the ImageNet Large Visual Recognition Challenge (ILSVRC2012). The computer vision crowd was looking forward to this event, which was held on the last day of the conference. Preliminary results had filtered out which reported an improvement by this year’s challenge winner of 40% over the second contender. The SuperVision method, by University of Toronto’s Alex Krizhevsky, takes a radical, groundbreaking approach to image classification: in short, they do not try to build high-level image descriptors as a preliminary to training. Instead, they rely entirely on a deep neural network to learn both the most appropriate linear features based on simple, low-level image patches and the classifier itself. Just like our brain does! This feat is made possible by the great size of the ImageNet dataset (more than one million images, which prevents the model from overfitting) and by a GPU implementation of back-propagation (the neural network training method).
All in all, a great week! We’ll definitely be back in two years, at the 2014 Zürich edition.
Update : The videos of the conference are now online !
Two weeks ago LTU was in Nancy for the 10th Quaero CTC/Corpus and plenary meetings.
For those who are not familiar with it, Quaero is a european research project which aims is to provide technologies multimedia content indexation.
We have seen some nice scientific presentations there. Among other the work Jérôme Revaud did on logo retrieval is of specific interest.
Even if we do not have all the details about their method, the work that was done at INRIA TexMex for the retrieval of images on a dataset of 100 millions images looks impressive.
They have chosen to keep all the image descriptors and thus had to manage a huge index. They have put most of their effort on the ability to manage such a huge amount of data using the possibilities offered by Hadoop.
As their method looks totally different from ours we are curious to compare their results with the one we got on the same dataset.
The results obtained by INRIA/Lear on at the Multimedia Event Detection task of Trecvid 2012 were also impressive. They scored 2nd at the task and were able to detect compex actions like “bike Trick”, “play hide and seek” or “giving directions to a location” with good accuracy.
As usual the demo session was a good time to show our demos from our brand new LTU Demo Studio and get feedback from both experts and tech lovers.
But the peak moment of the meeting was the improvised cocktail on the famous Stanislas place. The Museum where the cocktail was supposed to stand was closed because of a power outage so the organizers decided to do the cocktail directly at the place. As often, improvisation added spice to the event and made it a convivial and memorable moment.
