# US 2019/0171970 A1 · Trading goods based on image processing for interest, emotion and affinity detection

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2019/0171970 A1
- Document ID: US20190171970A1
- Status: Published application
- Kind: A1
- Application number: 16/267,049
- Filed: 2019-02-04
- Published: 2019-06-06
- Earliest claimed: 2015-10-06
- Assignees: International Business Machines Corporation
- Applicants: International Business Machines Corporation
- Inventors: SILVIA CHRISTINA SARDELA BIANCHI; Victor Fernandes Cavalcante; Paul Marques Caldeira Junior; Sergio Varga
- Web page: https://patentgrep.com/patent/US20190171970A1

## CPC

- G06Q 30/06: Buying, selling or leasing transactions
- G06N 20/00: Machine learning
- G06Q 30/0601: Electronic shopping [e-shopping]
- G06Q 30/0613: {using intermediate agents}
- G06Q 30/0619: {Neutral agent}
- G06Q 30/0641: {utilising user interfaces specially adapted for shopping}
- G06Q 30/0615: {Anonymizing}
- G06Q 30/0617: {Representative agent}
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

A system for trading goods over a computer network based on image processing includes an image processor that analyzes a picture received from a user to build a description of entities in the picture and correlations among the entities, a description interpreter that analyzes the description to determine an underlying message in the picture and detect information regarding the user's interests, emotions and affinities to goods and services, wherein the information is stored with the user's profile, the user's list of items to trade, and the user's past transaction history, and a trade matching component that uses the user's profile and produces a ranked list of other users with items to trade, or intermediary trade steps, when another user initiates a new trading transaction.

## Claims

### Claim 1

A non-transitory program storage device readable by a computer, tangibly embodying a program of instructions executed by the computer to perform the method steps for trading on a computer network, the method comprising the steps of:
receiving from a first user a request to trade a first item;
retrieving a first list of users with intention to acquire the first item from the first user;
ranking users in the first list of users based on a user affinity/emotion score for each user, wherein the user affinity/emotion score is higher than a first predefined threshold;
retrieving a first list of items that the first user has intention to acquire;
ranking items in the first list of items based on an item affinity/emotion score for each item, wherein the item affinity/emotion score is higher than a second predefined threshold;
retrieving for each user in the first list of users a second list of items that each said user has to trade;
adding a second item in the second list of items, the user and affinity/emotion score associated with said second item to a trade list, if said second item is also in the first list of items and if the item affinity/emotion score of said second item is higher than the second predetermined threshold; and
ranking the items in the trade list based on the item affinity/emotion score of each item.

### Claim 2 (depends on 13)

The computer readable program storage device of claim 13, the method further comprising broadcasting the ranked list of items to each user in the trade list.

### Claim 3 (depends on 13)

The computer readable program storage device of claim 13, the method further comprising, if the second item in the second list of items is not in the first list of items, retrieving a second list of users with intention to acquire the second item;
ranking each user in the second list of users based on a user affinity/emotion score for each said user, wherein the user affinity/emotion score higher than the first threshold;
retrieving for each user in the second list of users a third list of items that each said user has to trade;
adding an item in the third list of items, the user and affinity/emotion score associated with said item to the trade list, if said item is also in the first list of items, and ranking the items in the trade list based on the affinity/emotion score of each item.

### Claim 4 (depends on 1)

The computer readable program storage device of claim 1, the method further comprising generating said second list of items that each said user has to trade by
analyzing a picture received from each said user to build a description of entities in the picture and correlations among the entities;
analyzing the description to determine an underlying message in the picture and detect information regarding the user's interests, emotions and affinities to goods and services, wherein the information is stored with a profile of the user, a list of items to trade of the user, and a past transaction history of the user;
using the detected information regarding the user's interests, emotions and affinities to goods and services and producing a ranked list of other users with items to trade, when another user initiates a new trading transaction.

### Claim 5 (depends on 4)

The computer readable program storage device of claim 4, the method further comprising storing in a database the information regarding the user's interests, emotions and affinities to goods and services, with the user's profile, the user's list of items to trade, and the user's past transaction history.

### Claim 6 (depends on 4)

The computer readable program storage device of claim 4, the method further comprising using a machine learning algorithm that detects the user's interests, emotions and affinities to goods and services, wherein the machine learning algorithm is trained using descriptions and interpretations of a set of pictures to produce an inference function that can analyze new pictures from user sources.

### Claim 7 (depends on 4)

The computer readable program storage device of claim 4, the method further comprising using natural language processing to analyze the description.

### Claim 8 (depends on 4)

The computer readable program storage device of claim 4, the method further comprising using computer vision algorithms to build the description of entities in the picture and correlations among the entities, wherein said computer vision algorithms include conditional random fields for image segmentation and label prediction, a convolution neural network for feature extraction, and a support vector machine for classification.

### Claim 9 (depends on 4)

The computer readable program storage device of claim 4, the method further using a text analyzer to extract and evaluate each word in the description created by the image processor, comparing each word with a dictionary, and evaluating and weighting each word based on a pre-defined set of categories that define the user's interests, emotions and affinities to goods and services
