Community

Blog Events Webinars Tutorials Forum
  1. Blog
  2. Events
  3. Webinars
  4. Tutorials
  5. Forum
Create Account
×
Community Blog Finding Similar Images with Alibaba Cloud Image Search

Finding Similar Images with Alibaba Cloud Image Search

Alibaba Cloud Image Search helps you to discover accurate image results in milliseconds. Image Search is applicable to online shopping platforms and website image sharing applications.

When was the last time you successfully logged in to your favorite e-commerce website and found a pair of shoes that you liked? You are not alone; in fact, 96% of Americans with internet access have made an online purchase once in their lives. Moreover, four in five netizens have shopped in last month alone. (Source: www.bigcommerce.com/blog/omni-channel-retail/)

And would this be possible if customers were not able to view the product that they intend to purchase? Probably not. In this blog, we will understand what image search is, why it is important for businesses, and then dive into Alibaba Cloud Image Search. Let us begin.

Classification of Image Search Technology

Organizations are always trying their best to ensure that customers have a pleasant user experience when they are visiting the business' website. This includes prompt website response time, detailed information, and relevant images. For an e-commerce website, the criticality and role of image search are even more significant. This is because the image displayed on the potential buyer's screen contributes to the buying decision-making.

When referring to Image Search Technology, organizations might be confused between the various types of Image Search Technologies available out there. Despite being very similar, they do have significant differences between them. Below are the types of Image Search Technologies and their brief description.

  • Image Search: A search engine executes an image search engine, where it takes a text-based query and tries to find the accurate visual match. To perform an accurate visual search, search engines require much more sophisticated processes than they do for traditional image search.
  • Reverse Image Search: It includes providing the CBIR system with a sample image upon which it will base its search. The sample image formulates a search query and is characterized by a lack of search terms. It eliminates the requirement for a user to provide keywords or terms that may return an accurate result.
  • Visual Search: This form of image search technology involves an agile scan of the visual environment for a particular object. Examples of visual search are - choosing the freshest piece of apple in a market, searching for an object amongst piles of things, or finding your friend in a large crowd. Several visual search paradigms use eye movement to measure the degree of attention given to stimuli.

How Can Image Search Technology Help Businesses?

A pleasant online experience is a combination of images, content, and loading time. As I have mentioned above, for an e-commerce business, images play even an essential role in converting a website or application login to a sale. Below are the six ways in which Image Search Technology can help businesses.

  1. Provides an Edge over Keyword-based Search: Visual search “reads” images to identify shapes, sizes, and colors. It can also successfully recognize texts to identify product and brand names. Moreover, it liberates a user from thinking and putting in the effort to find the words to describe a product that he or she wants.
  2. Allows Social Discovery: Accenture predicts that Generation Z will prefer shopping through social media. It also suggests that almost 70% of young customers will purchase directly through social networks. It is here that visual search will prove to be a breakthrough for businesses to make social sales possible.
  3. Smarter Online Merchandising: E-commerce merchandising typically involves each product consisting of several tags. These tags allow customers to filter attributes and permits recommendation algorithms to surface related products. Retailers can also choose to override these algorithms to push a particular product.
  4. Facilitates Spearfishing: When it comes to purchasing, spearfishing refers to a purchase incident where the customer is aware of what he or she wants to buy and wants the result instantly. A potential buyer can spearfish much faster with visual inputs, as it offers better context than a typical keyword search. It also can deliver a finite set of search results.
  5. Helps in Cross Selling and Recommendations: There might be instances where an e-commerce website might not have the exact product that a customer is looking for at a particular time. The website can still show related or similar items that might make a purchase possible. Additionally, it can also help consumers to picture how related products might complement it.
  6. Promotes Social Listening: Organizations are engaging in social listening more than ever before. Primarily, brands monitor major social media platforms to analyze their services and products. Using Image Search Technology, brand managers can gather collective feedback on their products or services.

Moving on, let us now look at how Alibaba Cloud contributes to Image Search Technology.

Alibaba Cloud Image Search

Images are more than pictorial representations for businesses. The uses of Image Search Technology, which I have listed in the preceding section, should by now, help you realize how it can help businesses enhance their brand image, revenue, and customer experience. Keeping these aspects in mind, Alibaba Cloud launched Image Search in February 2018.

Alibaba Cloud Image Search is an intelligent image service based on machine learning and deep learning. It integrates search and image recognition, allowing users to search through images across multiple industries and business scenarios. It aims to provide users identical or similar images.

Five Key Features of Alibaba Cloud Image Search

Similar to several other Alibaba Cloud products, Alibaba Cloud Image Search comes with a number of advantages for users across businesses.

  • Precision: The entire concept of a search based on images is to achieve a result, which is more accurate than a keyword text. Alibaba Cloud Image Search yields precise image results for all industry-leading search services.
  • Vast Volume Support: With Alibaba Cloud MaxCompute at its base, Alibaba Cloud Image Search supports over ten billion items. This ensures that Image Search caters to businesses belonging to a spectrum of industries, types, size, and geographies.
  • Quick Response: A crucial element to the entire image search mechanism is the ability to get search results within a second. Alibaba Cloud Image Search can respond to a vast number of queries in milliseconds.
  • Real-time Interface: With its real-time interface, you can conveniently add or remove data dynamically. This ensures that you can update your website's content as per the current trends and norms. It also ensures that you can update your inventory on time.
  • Customization: Different organizations have unique requirements. Factors such as organization type, target market, and market size collectively determine the image search requirement. Alibaba Cloud Image Search offers flexibility to cater to individual requirements of businesses.

Application Scenarios of Alibaba Cloud Image Search

Currently, Alibaba Cloud Image Search applies to the following scenarios.

Application Scenario 1 – Image Search for Online Shopping

To enhance the experience of searching products online, Alibaba Cloud has integrated Image Search into the Taobao app, in the form of the Pailitao (snap and buy) feature. Since Pailitao's release in 2014, the number of unique daily visitors has been continuously increasing. This number grew from several hundred to more than ten million per day. The growth implies vast adoption of search by image feature for online shopping by users. Contrary to traditional keyword search, Pailitao requires only a simple photograph of the product, without any textual product description. This makes the procedure of online shopping convenient and enhances user experience.

1

Application Scenario 2 – Image Search on Photo Websites

Statista believes that there are 2.53 billion smartphone users today. By 2020, this number will reach 2.87 billion. This vast number have seen the emergence of photo sharing apps and mobile-based application for almost all brands. These websites and apps store billions of images and photos. However, their search support remains restricted to simple keywords. Image Search helps businesses utilize APIs to set up an image search engine that supports the search of billions of photos on an image basis. This will enhance the efficiency and user experience of the search option.

2

Conclusion

With the modern-day customer wanting more than ever before, image search technology is helping businesses provide a much-enhanced purchasing experience. Search queries performed through keyword inputs not only leave a room for accurate results but also is a complex way of completing accurate searches. With reverse image service, all that a user has to do is upload a link or a relevant image.

Alibaba Cloud Image Search is Alibaba Cloud's latest offering as part of its image searching solutions based on its Big Data services. It offers users with accurate results within a short span of time and is customizable as per individual business requirements.

Trends Big Data Machine Learning Image Search Computer Vision
1 1 1
Share on

Read previous post:

Alibaba Deploys Alibaba Open Channel SSD for Next Generation Data Centers

Read next post:

Building the Simplest Blockchain with Ruby

Alibaba Clouder

2,599 posts | 762 followers

You may also like

Comments

Raja_KT February 18, 2019 at 6:03 am

Query by example is still in struggling stage :)

Alibaba Clouder

2,599 posts | 762 followers

Related Products

More Posts by Alibaba Clouder

See All