Artificial Intelligence 6 min read

CCL2022 Video Highlight Extraction Challenge Overview

The article describes the CCL2022 Video Highlight Extraction Challenge, a competition at the 21st China Conference on Computational Linguistics organized by Baidu, inviting participants worldwide to generate timestamped concise summaries of video segments, with registration details, eligibility, task description, example inputs/outputs, and evaluation metrics based on timing accuracy and ROUGE-L.

Baidu Geek Talk
Baidu Geek Talk
Baidu Geek Talk
CCL2022 Video Highlight Extraction Challenge Overview

CCL2022 (the 21st China Conference on Computational Linguistics) is a major NLP conference organized by the Chinese Information Processing Society of China. It focuses on intelligent computing and information processing of Chinese languages, providing a high‑level platform for academic and technical exchange.

The challenge task is to extract video highlights from the provided video, title, and subtitles. Video highlights are short sentences that summarize the theme of each segment, forming an outline that helps users quickly understand the video content.

Registration

Registration period: 2022.06.01 – 2022.09.25 23:59:59

Registration link: see the original article link at the end of the post.

Consultation QQ group: 297701224

Organizer: Baidu Content Strategy Department

Prize

Official honor certificates will be awarded to teams that achieve outstanding results in the evaluation.

Eligibility and Requirements

Open to the whole society, regardless of age, identity, or nationality. Individuals, universities, research institutions, enterprises, and startups may all register. Teams may consist of up to 5 members and can be cross‑institutional, but each participant may join only one team.

Task Description

Participants are given a video (URL, duration > 60 seconds), the video title, and OCR‑derived subtitles (each subtitle includes its timestamp). The required output is a list of all video highlights, each containing a start time and a concise summary.

Input Example

[
    [10, "嗨,大家好,欢迎大家收看本期星座"],
    [13, "我是主持人顾"],
    [14, "今天我们来说一下12星座中"],
    [17, "三个星座月份是怎么样的"],
    [22, "白羊座,五月,赚到大钱"],
    [25, "白羊座的人特别古灵精怪"],
    [27, "他们总是在想尽一切办法怎么才能挣到钱"],
    [31, "他们的运气很好,总是能抓住好时机"],
    [35, "金牛座,八月,被喜欢的人表白"],
    [38, "金牛座的人在感情上是被动的"],
    [41, "不过在八月份他们会很幸运的被别人给表白"],
    [45, "这是金牛座之前不敢做的事,他们都是在暗恋别人"],
    [49, "摩羯座12月让你做最幸福的人"],
    [53, "摩羯座的人在工作中变现的很融洽"],
    [56, "虽然他们也想做出一些改变"],
    [58, "却没有找出很好的办法"],
    [60, "但是他们会得到同事们的帮助"],
    [63, "找到了和同事之间更好的相处模式"],
    [66, "恋人之间也会被另一方宠爱着"],
    [70, "你是不是也是其中的一个星座呢"],
    [73, "本期的顾星座就到这里了"],
    [76, "谢谢大家的收看"],
    [76, "我们下期再见"]
] # The numbers indicate the timestamp in seconds

Output Example

[
    {
        "summary": "三个星座月份是怎么样的",
        "begin": 17
    },
    {
        "summary": "白羊座",
        "begin": 22
    },
    {
        "summary": "金牛座",
        "begin": 35
    },
    {
        "summary": "摩羯座",
        "begin": 49
    }
] # "begin" is the start time, "summary" is the highlight description

Evaluation Metrics

The task is evaluated on two dimensions:

Start‑time accuracy: Precision, Recall, F1

Summary quality: ROUGE‑L

evaluation metricsNLPVideo SummarizationCCL2022highlight extraction
Baidu Geek Talk
Written by

Baidu Geek Talk

Follow us to discover more Baidu tech insights.

0 followers
Reader feedback

How this landed with the community

login Sign in to like

Rate this article

Was this worth your time?

Sign in to rate
Discussion

0 Comments

Thoughtful readers leave field notes, pushback, and hard-won operational detail here.