Set up your ViSenze account

Get a ViSenze account

To explore and implement the ViSenze’s artificial intelligence powered e-commerce solutions, please request an account.

Your account will give you access to the ViSenze Dashboard, where you can manage your ViSenze account -- add or edit team members, create and update applications (index of images), try our different solutions, analyze user actions, and monitor usage statistics.

Set up your application

Our image recognition powered solutions work on a specific collection of images called an “application” that you can create on our Dashboard. In other words, an application represents a collection of images you create with ViSenze. For each application, ViSenze will create a unique index. After an application is created, you will be able to upload, update or remove images indexed, retrieve and control API access keys, and view usage reports.

Create an application

To create an application, please sign in to your ViSenze Dashboard Account to get started. After signing in, you can create a new application by clicking the Create Application button on the top left corner of the page.

create-app

Choose your image types

When creating a new application, you will be asked to fill in two fields - your application name and your preferred image types. We have tailored our algorithms to optimize the search performance for different types of images.

app-type

Access your App Key

Once an application is created, you can view and create new app keys to be used in our solutions. There is a single key that uniquely identifies the collection of images that would be used to retrieve results for each of the solutions. You can find your app key by selecting one particular application from the dropdown menu in the top-left corner and selecting the Integration page in the left navigation bar. By default, one app key will be generated for each application you create.

application-key

You are allowed to create additional keys (maximum 10).

Upload your datafeed

To implement our search and recommendation solutions you will first need to index your catalogue with us especially the product images and associated product information such as brand, title, price, product url and etc. You can upload your data in three different ways:

  • One time upload : Upload a simple CSV (comma-separated file) on our dashboard.
  • Recurring upload : Set up a script on your servers that uploads a compressed CSV file (gzip format) to our FTP server. (available upon request)
  • Use our Data API to insert, update or remove your data. You can consider this if you need more flexibility and real-time updates.

Please note that we don’t support upload of raw image files as a .zip file or folder. You need to provide us the images as URLs (under a field called im_url) in your .csv file or Data API. These image URLs must be accessible publicly so our servers can download them for processing.

Configure schema fields

We recommend you to configure your application’s schema on the Dashboard before uploading any data. A “schema” refers to the specific set of fields you will host with the collection of images. An example schema could contain the following fields - im_name, im_url, brand, price, title, category, description. You may filter your search results based on any of the fields in the schema of the images. You may also retrieve and display any of these fields in the to the users when you use our Solutions. Please refer to the schema fields requirement section for a detailed explanation.

If you are using our FTP upload or Data API, then it is mandatory to configure your schema before. Please note that the schema fields that are not defined prior to the upload will not be indexed even though they are presented in your datafeed or Data API.

The schema configuration is optional (though recommended) if you use our CSV upload function on ViSenze Dashboard. In this case, our system will automatically identify the fields in the CSV file you upload and show you a confirmation screen to set up your schema.

To configure the schema for an application, you can go to Dashboard Your Images -> Schema section, you may add, edit, and remove the fields to fit your needs. Currently, you can only modify the fields via the dashboard.

When you are editing the schema, please note the following:

  • Removing a field will also remove the corresponding field from the image collection, which is unrecoverable.
  • Updating a schema may take some time to propagate. Your app may be unavailable for a short time.

config-schema

Schema fields requirement

Before preparing your datafeed for upload and configuring the schema, please read through the requirement for the fields in your schema.

Mandatory fields

Mandatory fields for all types of apps

Field Name Description Example
im_name The file-name of the product image, often in .jpg format audrey-red-polka-dot-swing-dress-p41-15865_zoom.jpg
im_url The URL to access the product image in a web browser http://lindybopcouk.cdn.visualsoft.co.uk/images/audrey-red-polka-dot-swing-dress-p41-15865_zoom.jpg

Mandatory fields for e-commerce apps

in order to integrate with our widgets, the following schema are mandatory

Schema Type Searchable Description Example
product_url string true The url of the product page containing the image https://weardex.com/#/product-detail?productId=670412e3b547097a5c684c5ac25335e2f519140c
category string true The category of the product top, bottom, outerwear, jewelry, hat, bag, intimates, eyewear, watch, etc.
brand string true The brand of the product, which will appear underneath the product image Versace, Gucci, Vera Wang, etc.
title string true The title of the product, which will appear underneath the product image Here's some examples for products in different categories:
  • Shoe: Black Leather Multi Buckle Flat Sandals
  • Top: Cold-Shoulder Ruffle Top; Knit Blouse
  • Bottom: Midi Skirt; Super Skinny Jeans; Striped Wide-Leg Pants
  • Bag: Cross Body Bag; Tote Bag; Zip Front Square Backpack; Quilted Shoulder Bag; Clutch Bag
  • Outerwear: Full-zip Sweatshirt, Printed Fleece; Leather Bomber Jacket; Quilted Parka With Faux Fur Hood
  • Dress: Sleeveless Wrap Front Skater Dress
description string true A short description of the product Here's some examples for products in different categories:
  • Shoe: Sandals by Kendall & Kylie Leather with open toe design and pin-buckle fastening over a chunky flat sole
  • Bag: Faux-leather outer with fully lined front flap, gold-tone twist lock, zip top closure, and interior zip pocket
price float true The original retail price of your product $49.99

Recommended fields

This list of schemas is not mandatory but is strongly recommended for eCommerce apps.

Schema Description Example
gender Gender who would typically wear the product Male / men, Female / women
sub_category A second-level, more detailed category of the product
  • Shoe: sandals, loafers, flats, heels, boots, sneakers
  • Top: t-shirt, blouse, tank top
  • Bottom: jeans, shorts, capris, slacks
  • Dress: bodycon, gown, summer, maxi
  • Bag: clutch, backpack, messenger
  • Outerwear: leather, blazer, jacket, parka, hoodie
discounted_price The discounted price of your product $24.99

You are allowed to define your customized schema fields as well. Please refer to the following section for the requirement of the customized schema fields.

Schema formatting

You can add and edit up to 30 customized fields containing the metadata of your images, exclusive of im_name and im_url fields. When performing image searches, you can use filter your image search results by any of the fields marked "Searchable" on the Dashboard. You can also specify the fields you wish to retrieve results from while searching.

There are some format requirements for the schema fields' name, value and image:

  • Maximum length of im_name is 128 characters.
  • im_url should be a url string for each image.
  • The field names can contain only alphabets(A-Z, a-z), numbers(0-9), and underscore(_).
  • The field names are case sensitive.
  • The length of a customized field name must be within 32 characters.
  • The length of a customized field value must be within 5000 characters.
  • For optimal results, we recommend images with at least 100 pixels in both dimensions and larger than 800x800 pixels.
  • The supported image formats includes: .bmp, .dib, .jpeg, .jpg, .jpe, .jp2, .png, .webp, .pbm, .pgm, .ppm, .sr, .ras, .tiff, .tif, .gif

Supported schema types

When you configure the schema, you can choose one the of the four supported field types below:

Note: Text type fields can only be used for filtering and cannot be returned in the results. This is to minimize the latency of response from our servers when users are searching or finding similar products

Type Explanation
String Short text as a whole or a word
text Long texts of multiple words and tokens
int Integer number, e.g. -1, 0, 9, 100
float Floating point number, e.g. 0.99, 99.99

Search filter configuration

You can configure the list of fields that can be filtered by by turning on the "searchable" flag. If a field is marked as searchable in the ViSenze Dashboard, you will be able to filter your image search results by providing a value to filter for the field.

Each field type has a filtering syntax and behavior:

Type Explanation
string Metadata value must be either
  • exactly matched with the query value, e.g. Vintage Wingtips would not match vintage wingtips or vintage
  • matched with multiple exact query values separated with OR operator. e.g. Wingtips OR Longwing would retrieve results that are matched by either Wingtips or Longwing
  • double quoted if it contains special character, e.g. "Women's shoes" OR "Women's Clothes"would retrieve results that are matched by either Women's shoes or Women's Clothes
int Metadata value can be either:
  • exactly matched with the query value
  • matched with a ranged query minValue,maxValue, e.g. int value 1, 99, and 199 would match ranged query 0,199 but would not match ranged query 200,300
float Metadata value can be either
  • exactly matched with the query value
  • matched with a ranged query minValue,maxValue, e.g. float value 1.0, 99.99, and 199.99 would match ranged query 0.0,199.99 but would not match ranged query 200.0,300.0
text Metadata value will be indexed using full-text-search engine and supports fuzzy text matching

Upload your image data

We provide three methods for you to upload your image data.

  1. For one-time upload and quick testing, you can upload your image data by submitting a .csv file using the dashboard image uploading feature. You can start the data upload on Dashboard by following this step-by-step guide.
  2. If you intend to perform a regular sync but in a less real-time frequency (ex: once a day), you can index your images through FTP upload. FTP Upload is not recommended for trial users.
  3. Alternatively, if you want a more flexible or automated upload method, you can upload your image data by calling our Data API. Please refer to the Data API Actions Guide for usage of Data API actions.

CSV uploading from Dashboard

Prepare a "comma separated values" (.csv) file in which the first row of the file should contain the metadata field names of your images. An example file looks like:

datafeed-example

Up to 32 metadata fields for each image is allowed. The maximum file size is 500MB. Refer to the Configure schema fields for detailed instructions on how the metadata should be provided.

FTP upload

If you want to do a regular upload of your datafeed, please follow this procedure to setup the FTP upload.

FTP upload procedure

ftp-instruction

  1. To upload your datafeed to our FTP server for processing, you will need to first need to setup an FTP account with ViSenze. To do so, please contact your account manager or write to support@visenze.com. After your FTP account has been enabled, you get a username and login to ViSenze’s FTP servers.

  2. After the FTP account is setup, you can find your FTP access credentials on Dashboard, Your Images->Upload and selecting Setup FTP Upload. You should be able to use the account name and password shown under your FTP account to connect to the FTP account.

  3. Ensure the .csv data file as the same format for Dashboard uploading and save the data file as a gzip compressed csv file following the naming conversion described in the requirement section. Please note that only .gz extension is supported for now. Other extensions such as .zip, .gzip etc are not supported.

  4. Upload an empty .txt file of the same format. Please note that without the empty .txt file, we cannot process the uploaded file automatically. The data file and empty .txt file should follow the same naming convention described below.

Supported modes

Two modes are supported for FTP upload:

Mode Description File name example
FULL (Default) the contents of this file will be used to fully replace your current image database. 0ea7xxxxxxxxxxxxxxxxxxxxxx.csv.gz or 0ea7xxxxxxxxxxxxxxxxxxxxxx_FULL.csv.gz
INCREMENT the contents of the datafeed file will be used to append new entries or update existing entries. If an delete file is provided, batch remove based on the data in the delete file will be performed after the successful index of the datafeed. 0ea7xxxxxxxxxxxxxxxxxxxxx_INCREMENT.csv.gz, 0ea7xxxxxxxxxxxxxxxxxxxxx_DELETE.csv.gz(optional)

Checking status

After the data file has been uploaded to the FTP server, it will be processed within five minutes. The indexing process can then be checked on Dashboard, Your Images -> Status

upload-status

DATA API

We provide a Data API for data indexing in a very flexible way. If you want to manage your image indexing efficiently (e.g to support fast moving fashion, flash sales, very quick updates to your database every minute, etc.), you can enhance your image indexing process by integrating with our Data API.

Using the API, you can perform the following actions on your product feed:

  • Insert (index / add) a product
  • Update a product
  • Remove (delete) a product
  • Check insert status

To learn how to perform these actions, refer to the Data API Actions Guide.

Integrate our solutions

Solutions overview

Now we have launched four solutions that would be fit into your various use cases.

  • Find similar: Surface visually similar products to users browsing your website. By looking to match shapes, styles, patterns and colors, our technology showcases products based on relevancy.

  • You may also like: Showcase additional products that your shoppers may also like based on relevancy to a product they are currently viewing

  • Search by image: Allow users to search by uploading images instead of guessing at keywords. With built-­in automated object recognition, our visual search will return matching or similar items from your database in the blink of an eye.

  • Search by color: Shoppers looking for a particular color item can directly search for all products in that color.

Start building

We provide easy-to-use tools for developers to integrate with our solutions. Please visit the overview page of each solution to start the actual integration:

Implement ViSenze analytics

The analytics reports page is located in your ViSenze Dashboard and allows you to conveniently see a high-level view of the performance of each of the solutions through click metrics as well as usage statistics. You can also drill down into the details or download the raw data for integration into other analyses.

Why implement this

With the analytics reporting tool, you can easily track the performance to quantify the value-add of our solutions through metrics such as click-through rate, add-to-wishlist rate, add-to-cart rate, and click rank. You can also quickly see and track changes in these metrics if you tweak your user-facing interface or as algorithm updates are released.

How to implement

Here is the list of actions that our analytics reports support:

Action recorded Description
click When user click to view the detail of the product
add_to_cart When user click the button to add the product to cart
add_to_wishlist When user click the button to add the product to wishlist

By integrating with our solution widgets, the implementation of tracking all these actions are already included.

However, if you would like to track actions being performed outside of the widgets UI or you want to customize your own UI, the action tracking needs to be implemented explicitly. Please refer to the guide for the integration with our SDKs:

Visualize the data on dashboard

To visualize the tracking statistics, you can go to dashboard and select the app you are integrating with, then go to Report / Analytics section to view and download the data.

tracking

Next steps

If you have any questions about setting up your account or integrating with our solutions, please refer to the FAQ Section or contact support support@visenze.com.

If you require extensive customization of your application, you may also contact our support team to work with you.