In english-see below.

Om virksomheden

Boogiz ApS er en virksomhed, der udvikler en original, interaktiv applikation, som inspirerer børn og unge til at danse. Boogiz v.1 findes på App Store siden august 2022.  

Om Boogiz

Boogiz er en hybrid mellem en spil og social platform, udviklet Unity og React Native.

I Boogiz kan brugeren ved hjælp af en avatar sammensætte sin egen danserutine og tilføje selvvalgt musik. Brugeren kan tilsvarende TikTok filme sig selv danse rutinen og dele denne film på Boogiz platform.

Overordnet om din opgave

Boogiz tilføres fra foråret 2023 gameplay/gamification, herunder blandt andet konceptet "Boogiz battle".

I "Boogiz battle" starter brugeren med at erhverver sig danse skills (de lære enkelte dansetrin, tilsvarende Fortnite dansetrin).

Når brugeren har ex. 3 danseskills kan denne fortsætte i spillet, og sammensætte en dans med sine skills. Brugeren for points for sine danse og kan anmode om at battle andre brugere på Boogiz sociale platform mm. mm.

Se forklarende video https://www.youtube.com/shorts/ILbzmc8bAzU

Læs gerne uddybende beskrivelse nedenfor.

Mere specifikt om din opgave:

Ønsket er at tilfører Boogiz app'en en scanning funktion, således brugerne kan se hvor gode de er til at efterligne avtarens bevægelser.

Det har en stor værdi for spillet, at brugerne får flere points når de gør sig umage med at lære trinnet. Det er blandt andet derfor at vi ønsker at supplere Boogiz gameplay med en scanning funktion.

Eksisterende tanker omkring tekniske muligheder

1. Scanning består af at brugerens dansebevægelser skal afkodes og omsættes til et digitalt format der lader sig sammenligne med avatarens. Udvikling af denne ”scanningsfunktion” foregår i softwaremotoren Unity, som også er den motor Boogiz er udviklet i. Følgende OpenPose er et bud på scanningsfunktion: https://github.com/CMU-Perceptual-Computing-Lab/openpose 

2. Optimering af scanning funktion. Brugerne skal kunne bruge Boogiz som mobil app spil, og uden brug af eksterne devises. Udfordringen bliver at få scanning til at fungere tilstrækkeligt på telefoner uden kraftige grafikkort (GPU). For at løse dette anbefales disse ”OpenPose” https://github.com/lujintaozju/OpenPoseForMobile samt https://github.com/Daniil-Osokin/lightweight-human-pose-estimation.pytorch 

3. Kontakt os for yderligere info.

Uddybende beskrivelse: Boogiz Battle forklaret i step 1-5  

1. Brugeren vælger et avatar dansetrin, øver sig på dette og får points for hvor dygtig denne er til at udføre trinnet. Trinet bliver nu ”aktiveret”, og transformeret til et ”battlekort”. Battlekortet findes i brugerens samling med et givent antal points. Denne proces er tilsvarende Fortnite dans, hvor brugeren indøver enkelte dansetrin og Just Dance hvor der gives points for at efterligne trinnet 

2. Brugerne kan se hinandens samling af ”battlekort”, hvilket ansporer til at øge sin egen samling af kort og points.

3. Brugeren ønsker at bruge sine ”battlekort” og vise sine kundskaber. Derfor kan de deltage i BB ved at sammen -sætte danse med sine ”battlekort”, pointscoren i battledansen er summen af de enkelte battlekort. 

   2+3) Dette svarer til ”Pokemon Go”, hvor man ”går” for at fange pokemons til sin samling, som man kan bruge til at battle andre.    

4. Brugeren øver sin battledans, filmer sig selv og anmoder om at deltage i BB.  

5. Boogiz algoritme finder brugeren en konkurrent der har tilsvarende antal points i sin dans, og opsætter en dansebattle ved at duette de to brugeres film. Duetten deles på delingsplatformen og vinderen af battlen er ex. den film der først når 1000 likes (duetter og måling af popularitet er allerede succesfuldt benyttet på ex TikTok).  

   
   

I Boogiz kan man udelukkende få positiv feedback, 100 % mobbefrit miljø.

In english:

About the company

Boogiz ApS is a company that develops original, interactive applications that inspire children and young people to dance. Boogiz v.1 is available on the App Store since August 2022.

About Boogiz

Boogiz is a hybrid between a game and social platform, developed Unity and React Native. In Boogiz, the user can use an avatar to put together his own dance routine and add music of his own choosing. Similarly to TikTok, the user can film himself dancing the routine and share this film on the Boogiz platform.

About your job

Boogiz will be added to gameplay/gamification from spring 2023, including the "Boogiz battle" concept. We need your help with "Boogiz battle".

In "Boogiz battle", the user starts by acquiring dance skills (they learn individual dance steps, corresponding to Fortnite dance steps). When the user has e.g. 3 dance skills, she can continue in the game and put together a dance with her skills and get points. Likewise, the user can request to battle other users on the Boogiz social platform, etc. See explanatory video https://www.youtube.com/shorts/ILbzmc8bAzU

Please read the detailed description below.

More specifically about your primary task:

The wish is to add a scanning function to the Boogiz app so that users can see how good they are at dancing and imitating the avatar's movements. The concept of "getting points for imitating the movement of the avatar" is equivalent to the game "Just Dance". It has great value that users can "make an effort" and get many points for their skills. That is why we want a scanning function where users can really see how good they are at doing the step.

Existing thoughts about technical possibilities

1. Scanning consists of the user's dance movements being decoded and converted into a digital format that can be compared with the avatar's. Development of this "scanning function" takes place in the software engine Unity, which is also the engine Boogiz is developed in. The following OpenPose is a bid for scanning function: https://github.com/CMU-Perceptual-Computing-Lab/openpose

2. Optimizing the scanning function. Users must be able to use Boogiz as a mobile app game, and without the use of external devises. The challenge will be to get scanning to work adequately on phones without powerful graphics cards (GPU). To solve this, these "OpenPose" https://github.com/lujintaozju/OpenPoseForMobile and https://github.com/Daniil-Osokin/lightweight-human-pose-estimation.pytorch are recommended

   
   

Detailed description: Boogiz Battle explained in steps 1-5

1. The user chooses an avatar dance step, practices this and gets points for how skilled he is at performing the step. The step is now "activated" and transformed into a "battle card". The battle card is found in the user's collection with a given number of points. This process is equivalent to Fortnite dance, where the user practices individual dance steps and Just Dance, where points are awarded for imitating the step

2. Users can see each other's collection of "battle cards", which encourages them to increase their own collection of cards and points. This is similar to "Pokemon Go", where you "go" to catch pokemons for your collection.

3. The user wants to use his "battle cards" and show his skills. Therefore, they can participate in BB by putting together dances with their "battle cards", the point score in the battle dance is the sum of the individual battle cards.

4. The user practices his battle dance, films himself and requests to participate in BB. 5) Boogiz's algorithm finds the user a competitor who has a similar number of points in his dance, and sets up a dance battle by duetting the two users' films. The duet shares on the sharing platform and the winner of the battle is ex. the film that first reaches 1000 likes (duets and measurement of popularity have already been successfully used on ex TikTok). I Boogiz you can only get positive feedback, 100% bullying free environment.